Session 4: Target and Agent Selection

Home , Abrostola, Abrostola asclepiadis, Actrix (moth), Altica rosae, Amsonia, Anthelephila

Session 4 Target and Agent Selection 123

Biological Control of Senecio madagascariensis (fireweed) in Australia – a Long-Shot Target Driven by Community Support and Political Will

A. Sheppard1, T. Olckers2, R. McFadyen3, L. Morin1, M. Ramadan4 and B. Sindel5

1CSIRO Ecosystem Sciences, GPO Box 1700, Canberra, ACT 2601, Australia [email protected] [email protected] 2University of KwaZulu-Natal, Faculty of Science & Agriculture, Private Bag X01, Scottsville 3209, South Africa [email protected] 3PO Box 88, Mt Ommaney Qld 4074, Australia [email protected] 4State of Hawaii Department of Agriculture, Plant Pest Control Branch, 1428 South King Street, Honolulu, HIUSA [email protected] 5School of Environmental and Rural Science, University of New England, Armidale NSW 2351 Australia [email protected]

Abstract

Fireweed (Senecio madagascariensis Poir.) biological control has a chequered history in Australia with little to show after 20 plus years. Plagued by local impacts, sporadic funding, a poor understanding of its genetics and its origins, and several almost genetically compatible native species, the fireweed biological control program has been faced with numerous hurdles. Hope has risen again, however, in recent years through the staunch support of a very proactive team of local stakeholders and their good fortune of finding themselves in a key electorate. The Australian Department of Agriculture, Fisheries and Forestry has recently funded an extendable two year project for exploration in the undisputed native range of fireweed in South Africa and a detailed search for agents that are deemed to be both effective and unable to attack closely related Australian Senecio species. This will be a tall order, but nevertheless is essential to conclude once and for all whether biological control has the potential to reduce the negative effects of the plant in south eastern Australian grazing and dairy country.

Introduction Hawaii (Ramadan et al., 2010). The plant contains genotoxic carcinogens in the form of pyrrolizidine Fireweed (Senecio madagascariensis Poir.) alkaloids which cause cumulative chronic liver is a toxic, short-lived, perennial, temperate to damage and fatality, especially in monogastric subtropical climate pasture weed of South African livestock like horses (Sindel et al., 1998; Bega Valley origin that has established and spread in Australia, Fireweed Association, 2008). Such toxins can enter the USA (Hawaii), Japan, Brazil, Argentina, the human food chain through nectar incorporated Venezuela, Columbia and Uruguay (Sindel et al., into honey or via dairy products from animals 1998). It is the focus of a biological control program grazing on infested land. The estimated control in Australia (Julien et al., 2012; Radford, 1997) and costs averaged around $9000 per farm per annum

XIII International Symposium on Biological Control of Weeds - 2011 124 Session 4 Target and Agent Selection

in Australia, and pasture productivity and profits international collaboration on this target weed. on infested land were reduced by between 15-50% Community support in Australia comes from an (Bega Valley Fireweed Association, 2008). Owners active landholder group in Bega, in south-east New of infested land also are at risk of suffering from South Wales (the Bega Valley Fireweed Association; reduced morale and stress-associated social costs www.fireweed.org.au), which is fortunate to have via; a) perception of declining property values, b) found itself in a key electorate at both the state and increased potential for neighbourhood conflicts, federal level. The federal Member of Parliament for c) regulatory obligations around weed control, d) this electorate has since become the Parliamentary animal health issues and costs, and e) longer-term Secretary for Agriculture, and has been highly business viability and succession issues. There also supportive of the fireweed issue, which has led to remains significant debate about whether and when fireweed biological control gaining funding in the the weed can be controlled using traditional pasture last few years and to the recent successful nomination management approaches, such as herbicides and of fireweed as a Weed of National Significance in controlled grazing. A high number of land owners, Australia. including commercial grazing businesses, have What makes this plant a difficult target for weed found such interventions to be impractical and the biological control in Australia is the taxonomic close price of chemicals too costly. proximity of fireweed to a group of Australian native In Australia, fireweed is spreading from a core species in the Senecio pinnatifolious (= S. lautus) group coastal strip in the south-east to the north and to more (Scott et al., 1998; Pelser et al., 2007). Hybridization upland and inland areas, and has certainly not yet can occur between these native species and fireweed, spread to all suitable regions of Australia, being still even if the progeny are sterile (Prentis et al., 2007). largely absent from the states of Victoria, Tasmania, Any natural enemies selected as potential biological South Australia and Western Australia (Sindel et control agents for Australia therefore need to be al., 2008). As a target for weed biological control, monospecific to S. madagascariensis. fireweed has many factors in its favour. Its taxonomy and native range are now well understood, being part of a complex of Senecio species in the KwaZulu-Natal Biological control history region of South Africa (Radford et al., 2000; Lafuma et al., 2003). It exhibits much lower abundance in Fireweed was declared a target for weed its native range, where competition from native biological control by the Australian Weeds perennial grasses in summer, largely confining the Committee in 1991. At the time, the taxonomy of plant to isolated pockets in disturbed fields, dunes the target was poorly understood and all the earlier and along roadsides (A. Sheppard, pers. obs.). While surveys for natural enemies were undertaken in a community of more than 30 natural enemies can south and eastern Madagascar (Marohasy 1989). be found on fireweed in Australia (Holtkamp and These early surveys recorded two moths; a flower- Hosking, 1993), none of these are very damaging feeding pyralid (Phycitodes sp.) and a stem-boring and none of the natural enemies from the native tortricid (Lobesia sp.), which were brought into range are already present in Australia, except for a quarantine in Australia. Both, however, failed to be rust fungus that appears to be of Australian origin specific enough for release in Australia (McFadyen anyway (Morin et al., 2009). Various other weeds of and Sparks, 1996). One 18- day visit was also made similar biology and impacts to fireweed have been to KwaZulu-Natal province in South Africa as part successfully controlled in Australia using classical of these activities in 1991 and identified 11 species of biological control, particularly Paterson’s curse insects feeding on the plant (Marohasy unpublished (Echium plantagineum L.) and ragwort (Jacobaea report - see Table 1 below). vulgaris Gaertn.) (Julien et al., 2012). Finally In 2002, Meat and Livestock Australia funded a fireweed, as a target for biological control, has short project to look at the potential of strains of rust very strong community support in both Australia on fireweed in South Africa as potential biological (Sindel et al., 2011) and Hawaii (M. Ramadan, pers. control agents. One such rust, Puccinia lagenophorae obs.) which has opened up new opportunities for Cooke, is considered to be native to Australia, but is

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 125

Table 1. Natural enemies found on S. madagascariensis in South Africa (KwaZulu-Natal) and Swaziland, South Africa.

Nature of attack Natural enemy type (family) Scientific name (current knowledge)

Stem-borers Weevil (Curculionidae) Gasteroclisus tricostalis (Thunberg) Moth (Tortricidae) Lobesia sp. Fly (Tephritidae) Coelopacidia strigata Bezzi Fly (Agromyzidae) Melanagromyza spp. Flower feeder Moth (Pyralidae) Homeosoma stenotea Hampson Moth (Pyralidae) Phycitodes sp. Moth (Pterophoridae) undetermined Fly (Tephritidae) Cryptophorellia peringueyi (Bezzi) Fly (Tephritidae) Trupanea inscia Munro Fly (Tephritidae) Sphenella austrina Munro Fly (Tephritidae) Telaletes ochraceus (Loew) Fly (Tephritidae) small undetermined Fly (Cecidomyiidae) undetermined Root feeders Weevil (Curculionidae) Proictes longehirtus Fairemaire Leaf beetle (Chrysomelidae) undetermined Sap suckers Leaf hopper (Tettigometridae) Hilda elegantula Gerstaecker Plant bug (Miridae) Ellenia sp.? Lace bug (Tingidae) undetermined Plant pathogens Yellow rust (Pucciniaceae) Puccinia lagenophorae Cooke & hybrids White rust (Albuginaceae) Albugo nr tragopogonis Flower smut (Ustilaginaceae) Ustilago sp. Stem blotch fungus undetermined

cosmopolitan in distribution on a range of Senecio lagenophorae rust isolates, or less virulent, in their species. It is also found in both Australia and South response to Australian fireweed plants (Morin et al., Africa on S. madagascariensis. Surveys in KwaZulu- 2009). Natal, South Africa were used to collected rust samples at 14 of 25 sites and accessions were tested from samples from eight of these sites (Morin et al., Biological control – future activities

2009). Taxonomic and genetic sequencing evidence confirmed that these accessions were a mixture The scientific case for a continuing biological of both P. lagenophorae and interspecific hybrids control program against fireweed in Australia has with P. lagenophorae as one of the parents. These been quite challenging, driven by the need for both rust accessions were found to attack Australian monospecific and effective agents. Although more fireweed plants and some failed to develop on the than 50–85% of national weed biological control native species tested, S. lautus subsp. lanceolatus, programs have led to significant or permanent weed in laboratory trials. However, all the South African control (Myers and Bazely, 2003), these constraints accessions were either comparable to Australian P. and the high associated chance of non-target impacts,

XIII International Symposium on Biological Control of Weeds - 2011 126 Session 4 Target and Agent Selection

suggests a very much lower chance of success. To be plant competition, climate and soil type. conservative, we have presented to the stakeholders that the chances of success might be in the order of b. Taxonomic and genetic variation in the 20%. The arguments in favour are as follows: target, through surveying the morphological variation of fireweed in KwaZulu-Natal a. There have been at least six weed biological and measuring the associated local genetic control programs in Australia where there variation. are native species in the same genus as the target weed. At least two of these have been c. The natural enemy community across successful to some degree; i.e. the programs sympatric Senecio species in South Africa, against Rumex pulcher L.and Jacobaea vul- looking specifically at host use across the garis Gaertn.. different Senecio species and natural enemy distribution and abundance. b. Monospecific agents are not hard to find in weed biological control programs. Most d. Experimental manipulations of both targets that are widespread in their native fireweed and the natural enemy densities range have at least some natural enemies to identify agents that are resource lim- that are monospecific. Also most plant ited and capable of causing high levels of pathogens used as biological control agents impact. Through this, fireweed growth and are monospecific and indeed many are reproductive effort, with and without natu- weed-genotype or biotype-specific pathot- ral enemies, will be measured. ypes. Researchers have also found biotype-

specific arthropod agents (e.g. two strains Based on surveys carried out to date and historical of Dactylopius opuntiae (Cockerell) on surveys made by Marohasy in 1991 and by Mohsen Opuntia ficus-indica (L.) Mill. or O. stricta Ramadan on more recent trips to South Africa for (Haw.) Haw. in South Africa). Hawaii, Table 1 shows the current list of known natural enemies. c. No in-depth studies have been completed in the native range of fireweed in South Conclusion Africa.

Fireweed biological control in Australia was Based on these arguments and on a workshop started nearly 20 years ago, but the amount of at the first National Fireweed Conference in Bega investment and effort has been quite limited to date. in 2008, the continuation of the fireweed biological Work has consisted of a few surveys to a poorly control program was supported by the stakeholders defined native range and the importation and basic and the federal Department of Agriculture, Fisheries testing of two moths and one pathogen, of which and Forestry. Funding was made available for this in only the latter was from the true native range. Two 2009. years ago, a strong local support group – the Bega The re-started biological control program is now Valley Fireweed Association – organised a National based in South Africa at the University of KwaZulu- Conference on fireweed and were lucky enough to Natal and will focus on studies of the following be heard and find themselves with some political issues: advantage at the 2007 federal election. This has a. The community and population ecology of led to some renewed federal funding for fireweed fireweed in its native range pastures, partic- biological control, despite the fact that the prospects ularly the roles of inter- and intra-specific for success are seriously constrained by the very close taxonomic affinity of fireweed to a group of native

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 127

Senecio species. Scientists argued that the chances Pelser, P.B., Nordenstam, B., Kadereit, J.W. & of success may be as little as 20%, but a lack of any Watson, L.E. (2007) An ITS phylogeny of tribe in-depth studies in the native range of the weed Senecioneae (Asteraceae) and a new delimitation suggested that continuation of the program may still of Senecio L. Taxon 56, 1077–1104. be worthwhile. The program has, therefore, recently Prentis, P., White, E., Radford, I.J., Lowe, A.J. & moved to the plant’s native range in South Africa to Clarke, A.R. (2007) Can hybridization cause local seek effective monospecific biological control agents extinction : The case for demographic swamping of for fireweed. the Australian native, Senecio pinnatifolius, by the invasive, S. madagascariensis? NewPhytologist, References 176, 902–912 Radford, I.J. (1997) Impact assessment for the biological control of Senecio madagascariensis Bega Valley Fireweed Association (2008) National Poir. (Fireweed). Thesis, University of Sydney. fireweed conference Bega 28th to 29th May 2008 Radford, I.J., Müller, P, Fiffer, S. & Michael, P.W. summary (www.fireweed.org.au), 25 pp. (2000) Genetic relationships between Australian Holtkamp, R.H. & Hosking, J.R. (1993) Insects and fireweed and South African and Madagascan diseases of fireweed Senecio madagascariensis, populations of Senecio madagascariensis Poir. and the closely related Senecio lautus complex. and closely related Senecio species. Australian In Proceedings of the 10th Australian and 14th Systematic Botany 13, 409–423. Asian-Pacific Weed Conference. pp 104–106. The Ramadan, M.M., Murai, K.T. & Johnson, T. (2010) Weed Society of Queensland, Brisbane. Host range of Secusio extensa (Lepidoptera: Julien, M., Cullen, M.J. & McFadyen, R.E. (2012) Arctiidae), and potential for biological control of Biological Control of Weeds in Australia. CSIRO Publishing, Melbourne (in press). Senecio madagascariensis (Asteraceae). Journal of Lafuma, L., Balkwill, K., Imbert, E.,Verlaque, R. & Applied Entomology 135, 269–284. Maurice, S. (2003) Ploidy level and origin of the Scott, L.J., Congdon, B.C. & Playford, J. (1998) European invasive weed Senecio inaequidens Molecular evidence that fireweed (Senecio (Asteraceae). Plant Systematics & Evolution 243, madagascariensis, Asteraceae) is of South African 59–72. origin. Plant Systematics & Evolution 213, 251– Marohasy, J.J. (1989) A survey of fireweed (Senecio 257. madagascariensis Poir.) and its natural enemies in Sindel, B.M., Radford, I.J., Holtkamp, R.H. & Madagascar with a view to biological control in Michael, P.W. (1998). The biology of Australian Australia. Plant Protection Quarterly 4, 139–140. weeds. 33. Senecio madagascariensis Poir. Plant McFadyen, R.E. & Sparks, D. (1996) Biological Protection Quarterly 13, 2–15. control of fireweed. In Proceedings of the 11th Sindel, B.M., Michael, P.W., McFadyen, R.E. & Australian Weeds Conference, (ed. Shepherd, Carthew, J. (2008). The continuing spread of R.C.H.) pp. 305–308. Weed Science Society of fireweed (Senecio madagascariensis) — the hottest Victoria Inc., Melbourne, Australia. of topics. In Proceedings of the 16th Australian Morin, L., van der Merwe, M., Hartley, D. & Müller, Weeds Conference (eds van Klinken, R.D., Osten, P. (2009) Putative natural hybrid between V.A., Panetta, F.D. & Scanlan, J.C.) pp. 47–49, The Puccinia lagenophorae and an unknown rust Weed Society of Queensland Inc., Brisbane. fungus on Senecio madagascariensis in Kwazulu- Sindel, B., Sheppard, A., Barnes, P. & Coleman, M. Natal, South Africa. Mycological Research 113, (2011) Flaming fireweed. In Proceedings of the 725–736. New South Wales Weed Society Conference, Myers, J.H. & Bazely, D. (2003) Ecology and Control Coffs Habour, NSW, Australia. 5 pp. of Introduced Plants. CUP, Cambridge, UK

XIII International Symposium on Biological Control of Weeds - 2011 128 Session 4 Target and Agent Selection

Prospects for the Biological Control of Tutsan (Hypericum androsaemum L.) in New Zealand

R. Groenteman1

1Landcare Research, PO Box 40, Lincoln 7640, New Zealand [email protected]

Abstract

The feasibility for biological control of tutsan, Hypericum androsaemum L., in New Zealand (NZ) was assessed. Conventional control methods are impractical and tutsan is not valued by any groups of society. It therefore makes a potentially good candidate for biological control. However, the lack of information about potential agents and the existence of four indigenous Hypericum spp. in NZ, including two endemics, are likely to prove challenging.

Introduction suitable habitat for this species. Tutsan is a garden escapee in NZ (Healy, Tutsan, Hypericum androsaemum L., is an 1972) and was first recorded as naturalised here in evergreen or semi-evergreen shrub (up to 1.5 m) of 1870 (Owen, 1997). The plant is well established the family Clusiaceae (alternatively Guttiferae). In throughout NZ (North and South Islands, Stewart New Zealand (NZ) tutsan has become a common Is, Chatham Islands, and Campbell Islands) (Sykes, weed in higher rainfall areas, growing in open 1982). It is currently of greatest concern in the forest, forest margins, scrub, waste places and Taumarunui District in the North Island of NZ. garden surroundings. Tutsan is shade tolerant, In NZ tutsan is considered a major pest in a range unpalatable to stock, and tends to infest areas in of bioclimatic zones from warm- to cool-temperate which mechanical and/or chemical control options (ranging from latitude 31° to 50° S, maritime climate, are impractical. below 600 m with average annual temperatures Tutsan’s extensive native range includes Europe, ranging between 12.5 and 22.5°C). Plant community Caucasia, Turkmenistan, Iran, Syria, Turkey, north- types identified as prone to invasion by tutsan west Africa and temperate Asia (Davis, 1967; USDA include shrublands, tussock grasslands and bare ARS, 2009). The naturalised range includes Australia, land. Tutsan can impact on the structure (i.e., on the NZ, Southern Africa, continental Chile and possibly dominant growth form of forest, shrubland etc.), or part of the US (Thomas, 2007). have a “major effect on many native species or on the A climate similar to that of southern France, composition or density of dominant species” (Owen, with average annual temperature of 13°C and 1997). annual rainfall of 910 mm, appears optimal A 1995 survey of weeds of conservation land for tutsan; however, tutsan can tolerate a wide determined its national distribution status as: temperature range (Van Der Veken et al., 2004). It is “established, widely distributed throughout NZ and also tolerant of various soil types and acidity levels extending its range into new habitats and areas”. (e.g., Hutchinson, 1967). Tutsan is a shade-tolerant Tutsan is a problem in regenerating forest (Sullivan species and, in its native range is a component of et al., 2007). Its biological success is mainly attributed mature forests (Olano et al., 2002). These findings to the high seeding ability per plant, seedbank suggest that large parts of NZ could prove to be persistence of >5 years, and its tolerance of semi-

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 129 shade conditions, hot or cold temperatures, high to confirming that tutsan is a sub-optimal host for moderate rainfall, damage and grazing. In addition, the beetles, and explaining why beetles released on its fleshy fruits are effectively dispersed by birds, and tutsan in the late 1940s quickly died out. possibly also by goats, possums, and soil and water movement (Whatman, 1967; Owen, 1997). Classical History of biological control of tutsan world- biological control is therefore a desirable option. wide The state of Victoria, Australia, initiated a In NZ there are four indigenous Hypericum spp. biological control programme against tutsan in the (Webb et al., 1988; Heenan, 2008, 2011): early 1990s. This programme was discontinued at an ● Hypericum involutum (Labill.) Choisy, na- early stage, prior to any surveys in the native range tive to NZ, Australia, Tasmania and New of the weed being carried out, after the rust fungus Caledonia Melampsora hypericorum (De Candolle) Winter was discovered to have self-introduced there. While the ● Hypericum pusillum Choisy, native to NZ, use of M. hypericorum as a biological control agent Australia and Tasmania. has generated mixed results, the fungus has largely successfully controlled tutsan in Victoria (Bruzzese ● Hypericum rubicundulum Heenan, en- and Pascoe, 1992; McLaren et al., 1997; Casonato et demic to the South Island of NZ (and al., 1999; David McLaren pers. comm.). known from one locality in the North Island) and considered naturally uncom- Objectives mon

● Hypericum minutiflorum Heenan, endem- Given the difficulties to control tutsan using con- ic to NZ, restricted to the central North ventional methods, and given it is rapidly expanding Island Volcanic Plateau and considered its range, classical biological control emerges as an nationally critically endangered attractive option. The objectives of the current study were, therefore, a) to review the literature to identify A high degree of host specificity would be required potential biocontrol agents for tutsan and assess the of any agent introduced against tutsan, if we were to feasibility of their release in NZ and, b) to assess the avoid significant non-target risks to the indigenous prospects of achieving successful biological control Hypericum species. There are no other indigenous of tutsan in NZ. representatives in the Clusiaceae family in NZ. History of biological control of tutsan in NZ Methods

Biological control of tutsan in NZ was attempted Identifying fungal pathogens of tutsan opportunistically in the late 1940s, using a St John’s wort biological control agent Chrysolina hyperici The information was obtained by searching (Forst.) (Julien and Griffiths, 1998). Adult beetles online databases and Internet sites. Online databases were observed feeding on tutsan, and subsequently, searched were: an attempt was made to release C. hyperici in USDA Fungus-host database or FDSM (which areas where tutsan was considered a problem. includes most NZ plant disease records): http:// Beetles released on tutsan between 1947 and 1950 nt.ars-grin.gov/fungaldatabases/fungushost/ all failed to establish on the weed (Miller, 1970). FungusHost.cfm Early instar larvae of both the lesser and greater Fungal Records Database of Britain and Ireland St. John’s wort beetles, C. hyperici and Chrysolina or FRDBI (Cooper, 2006): http://www.fieldmycology. quadrigemina (Suffrian) suffered high mortality net/FRDBI/assoc.asp when offered tutsan in recent no-choice laboratory IMI fungal herbarium (CABI Bioscience, 2006) feeding experiments, and the survivors’ development http://194.203.77.76/herbIMI/index.htm was severely impeded (Groenteman et al., 2011), NZ fungi and bacteria database or NZFUNGI

XIII International Symposium on Biological Control of Weeds - 2011 130 Session 4 Target and Agent Selection

(Landcare Research, 2009): http://nzfungi. Results landcareresearch.co.nz/html/mycology.asp; this database was also used to determine which species Extensive searches of the literature and were already present in NZ online databases yielded very few records of In addition, CAB abstracts, Current Contents, organisms attacking tutsan. This could reflect ISI Proceedings, Web of Science, Agricola, Science scarcity of herbivores and pathogens attacking Direct, Google and Google Scholar were searched, tutsan; but it could also reflect lack of interest using the terms “Hypericum androsaemum or in tutsan on behalf of entomologists and plant tutsan” and sub-searched using the terms “pathogen* pathologists, and consequently a potential or fung*”. Once a list had been created, further array of agents to discover. All but one of the information about each fungus was sought in the organisms recorded from tutsan were not specific published literature as well as in the following online to this species (see also Groenteman, 2009). databases: Index Fungorum database (Index Fungorum, Fungi 2004): http://www.indexfungorum.org/Names/ Names.asp Only 10 species of fungi have been reported in Global Biodiversity Information Facility or association with tutsan (Table 1). One was an GBIF (Global Biodiversity Information Facility, endophyte, which does not cause disease symptoms. 2009): http://data.gbif.org/species/ Five others could not be considered either because their host range is too broad or they are unlikely to be sufficiently damaging. Identifying arthropod biological control Four other pathogens may hold some potential agents for tutsan as biological control agents. The powdery mildew Erysiphe hyperici (Waller.) Fr. attacks various Unlike for fungal pathogens, comprehensive Hypericum species, and is troublesome for online databases for all arthropod herbivores do H. perforatum L. where the latter is cultivated for not exist. However, the following databases were its medicinal values (e.g., Radaitienë et al., 2002). It searched: may be worthwhile investigating whether a virulent For Lepidoptera, the Natural History Museum’s tutsan-specific strain exists. world listing (Natural History Museum London, Another powdery mildew, Leveillula 2007): http://www.nhm.ac.uk/jdsml/research- guttiferarum Golovin, has only been recorded from curation/research/projects/hostplants/ three Hypericum spp. That it has not been recorded Database of Insects and Their Food Plants from the highly studied H. perforatum suggests, Biological Records Centre (UK) (Biological Records perhaps, a relatively narrow host range. There is Centre (BRC), 2009) http://www.brc.ac.uk/dbif/ no information regarding the virulence of this Interpreting_foodplant_records.aspx pathogen and, its native range is not well matched Plant-SyNZ™ http://www.crop.cri.nz/home/ to NZ climate. plant-synz/database/hostplant.php The brown leaf spot Diploceras hypericinum In addition, CAB abstracts, Current Contents, (Ces.) Died. was recorded from tutsan in NZ ISI Proceedings, Web of Science, Agricola, Science and Japan, and in the Netherlands in the form Direct, Google and Google Scholar were searched of Pestalotia hypericina Ccs. It attacks other using the terms “Hypericum androsaemum or tutsan” Hypericum species and can cause severe dieback and sub-searched using the terms “invertebrate* or in H. perforatum. The virulence of this pathogen to herbivor*”. Checklists of NZ fauna were referred to, tutsan in NZ is not known, but could relatively easily to determine whether any of the species recorded be tested. In the Netherlands, conditions of nearly feeding on/infecting tutsan already occurs in NZ. 100% relative humidity were necessary to create

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 131 infection on tutsan in the laboratory (Van Kesteren, Four of these species are chrysomelid beetles, two 1963) so conditions for natural infection in the field of which, Chrysolina quadrigemina Suffrian and might rarely be met. Developing this pathogen into Chrysolina hyperici Forster, are well established a bioherbicide is an avenue that could potentially be in NZ and their performance on tutsan is poor. explored to overcome this limitation; however, this Chrysolina varians Schaller failed to establish in is an expensive pathway, unlikely to be economically Australia and North America as a biological control viable for tutsan. agent against H. perforatum (Currie and Garthside, Finally, the rust M. hypericorum was the most 1932; Currie and Fyfe, 1938; Coombs et al., 2004). common species recorded from tutsan, including San Vicente (2005) mentions tutsan and as host of in NZ. M. hypericorum was first recorded in NZ C. varians in Spain, yet does not explicitly treat H. in 1952 (Baker, 1955). It is unclear how the fungus perforatum as a host. Whether the Spanish C. varians has arrived here, and its effectiveness in controlling is a biotype adapted to tutsan is perhaps an avenue tutsan is variable (Baker, 1955; Whatman, 1967). to pursue. Lastly, Cryptocephalus moraei L. thrives M. hypericorum is also found in Australia, first on H. perforatum but not on tutsan (Tillyard, 1927). recorded in Victoria in 1991. By 1992 it had already shown phenomenal potential as a biocontrol agent Concluding remarks of tutsan (Bruzzese and Pascoe, 1992). Once a very common and invasive weed in south-western Available information about prospective Victoria, by 1997 tutsan was difficult to find in that biological control agents for tutsan is slim, and makes it region, resulting in “possibly the most spectacularly difficult to assess the prospects of successful biological successful example of weed biocontrol ever witnessed control at this time. However, it is clear that tutsan has in Victoria” (McLaren et al., 1997). never been the target of any extensive surveys, and Further attempts to use the rust as a biocontrol it is possible that a suite of potentially useful agents agent had mixed results: genetic variation between would be discovered should such a survey take place. tutsan populations suggested intrinsic resistance, and The genus Hypericum has four indigenous various rust isolates varied in virulence (Casonato et representatives in NZ, therefore highly specific agents al., 1999). are likely to be required. The findings from Australia highlight the Opposition to biological control of tutsan is importance of compatibility between genotypes unlikely. It is not grown here for medicinal purposes, and strains of fungal pathogens and their weedy nor is it highly valued as a garden plant. It is highly hosts, and suggest that as part of a biological unpalatable to stock and therefore not valued for control programme against tutsan in NZ it should fodder, nor is it valued for beekeeping. be determined what strains of tutsan and M. In a significant part of its range in NZ, tutsan hypericorum are present here and how they compare is a problem on terrain where mechanical and to those known from Australia. The hypothesis that chemical control methods are impractical. Therefore, observed variation in the impact of the rust against bioherbicides are not likely to be a practical (or tutsan is attributed to genetic variability of the weed, economic) solution and are not recommended as an the rust, or both should be examined. In addition, if avenue of future research for this weed. rust strains from Australia are absent from NZ, their virulence against NZ tutsan should be tested. Acknowledgements Arthropods I thank Hugh Gourlay for constructive discussions Only nine species of insects have been recorded and for comments; Stan Bellgard, Eric McKenzie, from tutsan, four of which can be immediately Seona Casonato and Dave McLaren for information precluded as potential agents due the breadth of and assistance; Tomas Easdale for translation from their host range (Table 2). Spanish and Alex Groenteman for translation from The remaining five insect species areDutch. This study was funded by the Sustainable oligophagous, but restricted to the genus Hypericum. Farming Fund, Contract no. 0809/93/014.

XIII International Symposium on Biological Control of Weeds - 2011 132 Session 4 Target and Agent Selection

References Groenteman, R., Fowler, S. V. & Sullivan, J. J. (2011) St. John’s wort beetles would not have been introduced to New Zealand now: A retrospective Baker, S. D. (1955) Note on tutsan rust in New host range test of New Zealand’s most successful Zealand. New Zealand Journal of Science and Technology Section A 36, 483–484. weed biocontrol agents. Biological Control 57, Biological Records Centre (BRC) (2009) Database of 50–-58. Insects and their Food Plants. Accessed: 05 Mar Healy, A. J. (1972) Weedy St. John’s worts (Hypericum 2009. spp.) in New Zealand. Proceedings of the New Bruzzese, E. & Pascoe, I. G. (1992) Melampsora Zealand Weed and Pest Control Conference 25, hypericorum, a rust fungus with potential in 180–190. the biological control of tutsan, Hypericum Heenan, P. (2011) Taxonomic notes on the New androsaemum. In Proceedings of the First Weed Zealand flora: Hypericum gramineum and Control Congress (eds J. H. Combellack & R. G. Hypericum involutum (Hypericaceae). New Richardson), pp. 101–102 Weed Science Society Zealand Journal of Botany 49, 133–139. of Victoria, Melbourne. Heenan, P. B. (2008) Three newly recognised species CABI Bioscience (2006) IMI fungal herbarium of Hypericum (Clusiaceae) from New Zealand. Accessed: 05 Mar 2009. New Zealand Journal of Botany 46, 547–558. Casonato, S. G., Lawrie, A. C. & McLaren, D. A. (1999) Hutchinson, T. C. (1967) Lime-chlorosis as a factor Biological control of Hypericum androsaemum in seedling establishment on calcareous soils. I. with Melampsora hypericorum. In Proceedings A comparative study of species from acidic and of the 12th Australian Weeds Conference (eds A. calcareous soils in their susceptibility to lime- C. Bishop, M. Boersma & C. D. Barnes), pp. 339– chlorosis. New Phytologist 66, 697–705. 342 Tasmanian Weed Society Publishing, Hobart, Index Fungorum (2004) World database of fungal Tasmania, Australia. names. Accessed: 05 Mar 2009. Coombs, E. M., Clark, J. K., Piper, G. L. & Julien, M. H. & Griffiths, M. W. (1998) Biological Cofrancesco Jr, A. F. (eds), (2004) Biological control of weeds: A world catalogue of agents and Control of Invasive Plants in the United States. their target weeds. CABI Publishing, Wallingford, Oregon State University Press, Corvallis, OR. 467 U.K. 223 Pp. pp. Landcare Research (2009) NZFUNGI - New Zealand Cooper, J. (2006) The fungal records database of Fungi (and Bacteria). Accessed: 07 Mar 2009. Britain and Ireland (FRDBI). Accessed: 05 Mar McLaren, D. A., Bruzzese, E. & Pascoe, I. G. (1997) 2009. The potential of fungal pathogens to control Currie, G. & Garthside, S. (1932) The Possibility of Hypericum species in Australia. Plant Protection the Entomological Control of St. John’s Wort in Quarterly 12, 81–83. Australia: Progress Report. Melbourne, Australia Miller, D. (1970) Biological Control of Weeds in 28 pp. New Zealand 1927–48. Department of Scientific Currie, G. A. & Fyfe, R. V. (1938) The fate of certain and Industrial Research, Wellington. European insects introduced into Australia for Natural History Museum London (2007) HOSTS - a the control of weeds. Journal of the Council for Database of the World’s Lepidopteran Hostplants. Scientific and Industrial Research 11, 289–301. Accessed: 07 Mar 2009. Davis, P. H. (ed) (1967) Flora of Turkey and the Olano, J. M., Caballero, I., Laskurain, N. A., Loidi, J. East Aegean Islands. Edinburgh University Press, & Escudero, A. (2002) Seed bank spatial pattern Edinburgh. in a temperate secondary forest. Journal of Global Biodiversity Information Facility (2009) Vegetation Science 13, 775–784. GBIF. Accessed: 07 Mar 2009. Owen, S. J. (1997) Ecological weeds on conservation Groenteman, R., (2009) Prospects for biological land in New Zealand: a database. Department of control of tutsan (Hypericum androsaemum L.). Conservation, Wellington, New Zealand. contract report, Landcare Research, Lincoln, New Radaitienë, D., Kacergius, A. & Radušiene, J. (2002) Zealand. 25 pp. Fungal diseases of Hypericum perforatum L. and

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 133

H. maculatum Crantz. in Lithuania. Biologija 1, USDA ARS (2009) GRIN - Germplasm Resources 35–37. Information Network. Accessed: 09 Apr 2009. San Vicente, I. U. (2005) Coleópteros fitófagos Van Der Veken, S., Bossuyt, B. & Hermy, M. (2004) (Insecta: Coleoptera) de los encinares cantábricos Climate gradients explain changes in plant de la Reserva de la Biosfera de Urdaibai. A. N. A. community composition of the forest understory: N. Elkartea Zapatari, Spain 201 pp. An extrapolation after climate warming. Belgian Sullivan, J. J., Williams, P. A. & Timmins, S. M. Journal of Botany 137, 55–69. (2007) Secondary forest succession differs through Van Kesteren, H. A. (1963) Leaf spot and bark naturalised gorse and native kanuka near Wellington necroses on Hypericum spp. associated and Nelson. New Zealand Journal of Ecology 31, with Festalotia hypericina. Verslagen van de 22–38. Plantenziektenkundige Dienst te Wageningen Sykes, W. R. (1982) Checklist of dicotyledons 138, 187–189. naturalised in New Zealand 12. Haloragales, Webb, C. J., Sykes, W. R., Garnock-Jones, P. J. & Myrtales, Proteales, Theales, Violales (excluding Given, D. R. (1988) Flora of New Zealand. Volume Violaceae). New Zealand Journal of Botany 20, IV, Naturalised Pteridophytes, Gymnosperms, 73–80. Dicotyledons. Botany Division D.S.I.R., Thomas, P. (2007) Global Compendium of Weeds Christchurch, N.Z. 1365 Pp. Accessed: 05 Mar 2009. Whatman, A. (1967) Tutsan - economic control Tillyard, R. J. (1927) Biological Control of St. John’s of a problem weed. New Zealand Journal of Wort. New Zealand Journal of Agriculture 35, Agriculture 115, 24-7. 42–45 pp.

XIII International Symposium on Biological Control of Weeds - 2011 134 Session 4 Target and Agent Selection - No / ? No since Yes, 1950s early / Offers con partial in some trol areas / No! Yes Found Found in New Zealand/ biocontrol potential? - re yet Not here corded / No exotic Yes, / No : 3 spp., spp.,

from H. scabrum 4 infect H. perforatum failed to strain ; one record in IF record ; one highly specific H. androsaemum (to Yes, the species been had that Note strains). Hypericum other various from recorded spe- woody plant unrelated many Attacks No. strictlyanymore, classified as a fungus cies. Not algae). (akin brown to stage life a mobile to due process ‘Key as a (classified invasive Highly in Australia’) biodiversity threatening - the H. an ; however, H. perforatum including drosaemum in Australia Likely be specific? host to comments) (and from hosts of a wide range from Recorded No. families plant various in FDSM records Two specific. genus Possibly - fami in multiple genera multiple Attacks No. lies - H. helian from , one one from H. androsemum themoides Yes. Has been highly Has Yes. - successful in control H. androsaemum ling Australia. in Victoria, in damaging Highly New Zealand. parts of - a highly aggres Yes, species sive Likely be- damag to ing? Does not Endophytic. disease- cause symp - in H. androsae toms mum - informa Insufficient tion - informa Insufficient tion 2 , and their potential usefulness for biological control. UK, Ireland, UK, Ireland, Scotland, Canada (BC), Australia, Japan, Bulgaria, New Zealand, USSR Japan Japan Iran Iran Range on H. Range on androsaemum - 1 ) (Lév.) (Lév.) Hypericum androsaemum Melampsora hypericorum Schröt. (DC.) J. Phytophthora cinnamomi Rands Guignardia endophyl & Nakagiri Okane, licola Ito Tad. Leveillula taurica G. Arnaud (= Erysiphe taurica guttiferarum Leveillula Golovin Species names) other (and

Basidiomycota Pucciniales Melampsoraceae Chromista Oomycota Pythiales Pythiaceae Ascomycota Botryosphaeriales Botryosphaeriaceae Ascomycota Erysiphales Erysiphaceae Ascomycota Erysiphales Erysiphaceae Phylum Order Family Table 1. Fungi recorded on tutsan, Table

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 135 - Found Found in New Zealand/ biocontrol potential? Yes/No Found in Found New Zea - as land - D. hyperici num No / No - / Possi No if a spe - bly, strain cific found is No / ? No since Yes, 1950s early / Offers con partial in some trol areas / No! Yes Found Found in New Zealand/ biocontrol potential? - re yet Not here corded / No exotic Yes, / No : 3 spp., spp., has 8 records 8 records has 5

from H. scabrum spp. Had been- col Had spp. 4 (strawberry) in plants - H. andro from none spp., spp. FDSM has 149 records from from 149 records has FDSM spp. infect H. perforatum failed to strain ; one record in IF record ; one and additional 91 from various other other various 91 from additional and Yes, at the genus level. FRDBI level. the genus at Yes, Likely be specific? host to comments) (and No. Attacks hosts from multiple families. multiple from hosts Attacks No. Attacks other Hypericum other Attacks ) P. hypericina Canada (as No, associated with plants from various various from plants with associated No, families Fragarialected from but 39 from H. perfo - 39 from but from H. androsaemum ratum Hypericum Hypericum various saemum highly specific H. androsaemum (to Yes, the species been had that Note strains). Hypericum other various from recorded spe - woody plant unrelated many Attacks No. strictlyanymore, classified as a fungus cies. Not algae). (akin brown to stage life a mobile to due process ‘Key as a (classified invasive Highly in Australia’) biodiversity threatening - the H. an ; however, H. perforatum including drosaemum in Australia Likely be specific? host to comments) (and from hosts of a wide range from Recorded No. families plant various in FDSM records Two specific. genus Possibly - fami in multiple genera multiple Attacks No. lies - H. helian from , one one from H. androsemum themoides 7 ). Brown leaf leaf ). Brown - requires 100% - requires Yes. Has been highly Has Yes. - successful in control H. androsaemum ling Australia. in Victoria, in damaging Highly New Zealand. parts of - a highly aggres Yes, species sive Yes, considered a considered Yes, disease of troublesome gets and H. perforatum fungicides with sprayed culti is - the latter where medicinal its for vated A powderyvalues. mildew Likely be damaging? to Likely be- damag to ing? Does not Endophytic. disease- cause symp - in H. androsae toms mum - informa Insufficient tion - informa Insufficient tion Probably not. Wood rot rot Wood not. Probably - decay dead and (attacks also live wood, but ing be to likely wood). Not very damaging. Causes leaf blight and and blight leaf Causes dieback stem severe . Not in H. perforatum - H. androsae to virulent mum to RH post-inoculation (in symptoms produce Pestalotia of the form hipericina spot No, saprobe 2 2 , and their potential usefulness for biological control. UK, Ireland, UK, Ireland, Scotland, Canada (BC), Australia, Japan, Bulgaria, New Zealand, USSR Japan England, Scot - England, land Japan Iran Iran Range on H. Range on androsaemum H. Range on androsaemum New Zealand Netherlands (as (as Netherlands ), P. hypericina New Zealand, Japan Ireland - - 1 1 ) (Lév.) (Lév.) Hypericum androsaemum (Pers.) Bres. (Pers.) Melampsora hypericorum Schröt. (DC.) J. Phytophthora cinnamomi Rands (Wallr.) (Wallr.) hyperici Erysiphe S. Blumer Guignardia endophyl & Nakagiri Okane, licola Ito Tad. Leveillula taurica G. Arnaud (= Erysiphe taurica guttiferarum Leveillula Golovin Species names) other (and Species names) other (and Hymenochaete cinnamo Hymenochaete mea Diploceras hypericinum hypericinum Diploceras (Ces.) Died. hypericina Ces. Pestalotia hypericinum Hyaloceras (Ces.) Sacc. - hyperici Seimatosporium num Sutton (Ces.) B. Melomastia mastoidea Schröt. J. (Fr.)

6 Basidiomycota Pucciniales Melampsoraceae Chromista Oomycota Pythiales Pythiaceae Ascomycota Ascomycota Erysiphales Erysiphaceae Ascomycota Botryosphaeriales Botryosphaeriaceae Ascomycota Erysiphales Erysiphaceae Ascomycota Erysiphales Erysiphaceae Phylum Order Family Phylum Order Family Basidiomycota Hymenochaetales Hymenochaetaceae Ascomycota Xylariales Amphisphaeriaceae - Xylari Ascomycota ales Table 1. Fungi recorded on tutsan, Table

XIII International Symposium on Biological Control of Weeds - 2011 136 Session 4 Target and Agent Selection Found Found in New Zealand/ biocontrol potential? Yes/No Possibly not. Veticillium spp. attack woody attack spp. Veticillium not. Possibly of A number families. plant various of hosts the Unwanted on listed are spp. Verticillium register. Organism http://www1.maf.govt.nz/uor/searchframe. htm Likely be specific? host to comments) (and are listed here. It may also be found elsewhere on other hosts. . If a fungus was listed under an out-of-date name (synonym) this is also stated in If . Insufficent information. information. Insufficent - Verticil Plant-pathogenic in various exists spp. lium in variation with strains range. virulence host and to cause known are They - in suscepti wilting severe symptoms no but hosts, ble hosts. in tolerant Likely be damaging? to H. androsaemum - 2 New Zealand Range on H. Range on androsae mum 1 sp. Nees Nees sp. Verticillium anam.] [stat. Species names) other (and

Ascomycota Hypocreales Monoliniaceae Phylum Order Family Many fungi have more than one Latin name because they can produce more than one type of spore. The name given when they are producing ‘sexual’ spores is are producing ‘sexual’ The name given when they spore. can produce more than one type of Many fungi have more than one Latin name because they Only the places where organism was found associated with Fungus-host database at http://nt.ars-grin.gov/fungaldatabases/fungushost/FungusHost.cfm FDSM = USDA database of fungal names at http://www.indexfungorum.org/Names/Names.asp IF = Index Fungorum, World FRDBI = the Fungal Records Database of Britain and Ireland (FRDBI) at http://194.203.77.76/fieldmycology/FRDBI/FRDBI.asp Insertae sedis = of uncertain taxonomic position within a higher order (e.g. Phylum known, but that phylum uncertain). An organism using dead organic material as food and commonly causing its decay (Kirk et al. 2001). Unlikely to cause disease and therefore probably Saprobe: called the teleomorph, whereas the stage producing ‘asexual’ spores is called the anamorph. The two stages often look completely 1 gives the taxonomy of the teleomorph, but column 2 uses whichever different. Fungi Table is the only are form known. So, name, unless the ‘anamorph’ classified according to their ‘teleomorph’ name/names were recorded when the fungus was found on H. androsaemum column 2. insufficiently damaging to be useful for biocontrol. 1 2 3 4 5 6 7

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 137 spp. Possibly more common common more Possibly spp.

Likely be host specific? sufficiently to H. perforatum. on No. Polyphagous. A pest of strawberry. A pest of Polyphagous. No. Calluna spp. and Erica spp. on Feeds No. Hypericum various on Feeds not. Possibly crops many of Pests polyphagous. Highly No. H. on common more and spp., Hypericum various on Feeds not. Possibly in Spain? biotype H. androsaemum an of Possibility perforatum. H. on Doesthrive not spp. Hypericum various on Feeds not. Possibly androsaemum H. androsaemum. on poorly Performs H. perforatum. Prefers No. New and Australia to (indigenous H. involutum Reported on feeding Zealand) in Australia H. androsaemum on poorly Performs H. perforatum. Prefers No. Range - Medi Asia, Northern Europe, Basin terranean (UK), Dutch Wight of Isle K. ericae) (as Isles Frisian West Palaearctic - Zealand Aus (endemic), New tralia (introduced) West to Spain (from Europe Siberia) Europe Africa Den to - North From - Austra to Introduced mark. Zealand North lia, and New America central northern and to Native Asia. western and Europe New Australia, to Introduced America. Zealand North and Hypericum androsaemum. Type of of Type organism Whitefly bug Ground Leaf mining moth Leafrollers Leaf beetle Leaf beetle Leaf beetle Leaf beetle -

)

Walker Walker Walker (Stainton) (= Fomoria Species fragariae Aleyrodes (=lonicerae) Kleidocerys truncatulus ericae (Walker) Ectoedemia septembrella herana Ctenopseustis and & Rogenhofer Felder C. obliquana Planotortrix excessana Octo Dug - P. and Walker dale varians Chrysolina (Schaller) Cryptocephalus (L.) moraei quadrigemina Chrysolina (Suffrian) (For hyperici Chrysolina ster) HETEROPTERA LEPIDOPTERA COLEOPTERA Order and Family and Order HEMIPTERA Aleyrodidae Lygaeidae Nepticulidae Tortricidae Chrysomelidae Table 2. Records of invertebrates feeding on tutsan Table

XIII International Symposium on Biological Control of Weeds - 2011 138 Session 4 Target and Agent Selection

The Use of Ascochyta caulina Phytotoxins for the Control of Common Ragweed

M. Cristofaro1, F. Lecce2, F. Di Cristina2, A. Paolini2, M. C. Zonno3, A. Boari3 and M. Vurro3

1ENEA C.R. Casaccia UTAGRI-ECO, Via Anguillarese, 301 00123 S. Maria di Galeria (Rome), Italy, [email protected] 2BBCA- Biotechnology and Biological Control Agency, Via del Bosco, 10 00060 Sacrofano (Rome), Italy, [email protected] 3Institute of Sciences of Food Production, CNR, Via Amendola 122/O, 70125, Bari, Italy, [email protected]

Abstract

Common ragweed, Ambrosia artemisiifolia Bess. is an alien weed of North American origin that became invasive in Europe. Although the main concern regarding this plant is its impact on human health, due to abundant production and easy spread of its highly allergenic pollen, A. artemisiifolia is also increasingly becoming a major weed in agriculture, being very competitive with crops and difficult to manage. Previous research had lead to the identification of a mixture of three toxins from the culture filtrates of Ascochyta caulina (P. Karst.) Aa & Kestern, a fungal pathogen of Chenopodium album L., having high toxicity against both host and non-host plant leaves, with lack of antibiotic and zootoxic activities. Recently, a two-year research project named ECOVIA has been approved and financially supported by the Regional Governorate of Lombardy (Italy). The main aim of the project is to develop the technologies to obtain a natural herbicide based on the bioactive toxins produced by A. caulina, and to study the possibility of its practical use. Among the selected target weeds, A. artemisiifolia showed an evident response to the toxin application, probably due to the presence of trichomes on the leaves that allow a quick absorption of the metabolites, with fast appearance of leaf necrosis and plant desiccation. Greenhouse bioassays as well as open field experiments clearly indicate the lethal effects on the young ragweed plants.

Introduction Previous research had lead to identify a mixture of three toxins from the culture filtrates of Ascochyta caulina (P. Karst.) Aa & Kestern, a fungal pathogen Common ragweed, Ambrosia artemisiifolia of Chenopodium album L., showing a wide range of Bess. is an alien weed of North American origin phytotoxicity against different weed species. that became invasive in Europe. Although the main Recently, a two-year research project named concern regarding this plant is its impact on human the ECOVIA has been approved and financially health, due to abundant production and easy spread supported by the Regional Governorate of Lombardy of its highly allergenic pollen, A. artemisiifolia is also (Italy). The main aim of the project is to develop increasingly becoming a major weed in agriculture, technologies for obtaining a natural herbicide based being very competitive with crops and difficult to on the bioactive toxins produced by A. caulina, and manage. to study the potential of its practical use.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 139

During 2010 and 2011, in the framework of weed is likely due to the presence of high numbers the ECOVIA research project, among the 16 weed of trichomes on the leaves (Grangeot et al. 2006) species we tested, A. artemisiifolia was evaluated that allow a quick absorption of the metabolites, both in laboratory and semi-field conditions. with the fast appearance of leaf necrosis and plant desiccation. We obtained a significant damage with cut leaves treated with the 2 mg/ml solution in Petri Materials and Methods dishes (Fig. 3). The same solution (2 mg/ml) caused a leaf damage of less than 50% when sprayed on the Toxin production plantlets, but its phytotoxicity has been significantly improved using the toxin in combination with the The fungus, A. caulina, a biological control agent Codacide®. of the weed C. album, was grown on a liquid defined Results presented here show promise that these mineral medium (named M1-D) for 4 weeks in static toxins can be effective in controlling common conditions as previously described (Evidente et. al., ragweed. Bioassays are in progress to further develop 1998).Toxins were purified as described (Evidente methods of application of the toxin mixture in order et al., 1998; 2000) and obtained as a pure mixture to achieve higher efficacy as a natural herbicide for containing three main phytotoxic compounds (i.e., the control of common ragweed. ascaulitoxin, its aglycone, and trans-4-aminoproline; Fig. 1). Acknowledgements

We thank A. Evidente (University of Naples, Laboratory and screen house bioassays Italy) for chemical extraction of fungal toxins, F. Vidotto (University of Turin, Italy) for providing Young potted ragweed plantlets were sprayed us ragweed seeds, L. Smith and T. Widmer (USDA- uniformly with an 8 mg/ml water solution of the ARS) for reviewing the paper draft, and L.M. purified fungal toxins by uniform nebulisation Padovani, P. Carrabba and C. Tronci (ENEA C.R. of leaf surface. A relative humidity of 100% was Casaccia, Rome, Italy) for their support in ECOVIA maintained for 24 hours after each application. The project. experiment was successively repeated with a lower concentration of toxins solution (2 mg/ml) on cut References leaves in Petri dishes and on potted plantlets. In order to enhance the efficacy of the toxins, further bioassays were performed by adding a Evidente, A., Capasso, R., Cutignano, A, Taglialatela- wetting agent to the toxin solution. The following Scafati, O., Vurro, M., Zonno, M.C. & Motta, four commercial products, which appear to best fit A. (1998) Ascaulitoxin, a phytotoxic bis-amino our needs, were used at the labelled concentration: acid N-glucoside from Ascochyta caulina. Biopower®, Adigor®, Etravon® and Codacide®. Phytochemistry 48, 1131-1137. Evidente, A., Andolfi, A., Vurro, M., Zonno, M.C. Motta A. (2000) Trans-4-aminoproline, a Results phytotoxic metabolite with herbicidal activity produced by Ascochyta caulina. Phytochemistry A. artemisiifolia showed an evident response 53, 231-237. to the toxin application. In particular, the 8 mg/ml Grangeot, M., Chauvel, B. & Gauvrit, C., 2006. solution showed a high herbicide effect on young Spray retention, foliar uptake and translocation plantlets with 100% leaf damage within 3-5 days after of glufosinate and glyphosate in Ambrosia the application (Fig. 2). The sensitivity of the target artemisiifolia. Weed Research, Vol. 46, 2 152-162.

XIII International Symposium on Biological Control of Weeds - 2011 140 Session 4 Target and Agent Selection

H OH COH H CH 2 OH H COH O CH CH 2 CH CH CH 2 CH COH HO CH 2 H O CH CH CH CH CH CH COH 2 NH OH 2 HOHO NH 2 NH 2 H OH NH OH HO NH 2 NH 2 H OH H H H H ascaulitoxin

HO C CH CH 2 CH CH CH 2 CH CO OH CH CH CH CH CO OH HO C CH CH 2 2 NH 2 NH 2 OH NH 2 NH OH NH 2 2 NH 2

ascaulitoxin aglycone H H CO OH The culture filtrate can be purified by CO OH using a simple and cheap method, NH 2 NH obtaining a toxic hydrophilic and 2 N N H almost pure mixture of the three H H metabolites as a yellowish powder. H trans-4-aminoproline

Figure 1. The toxin mixture is a yellowish powder containing three compounds: a caulitoxin, ascaulitoxin aglycone, and trans-4-aminoproline.

B C

Figure 2. Phytotoxicity of an 8 mg/ml toxin solution on ragweed plantlets: A) treated plants (left) and control plants (right); B) treated plant, C) control plant.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 141

Figure 3. Effects of a 2 mg/ml toxin solution on ragweed cut leaves (control - left, treated - right).

XIII International Symposium on Biological Control of Weeds - 2011 142 Session 4 Target and Agent Selection

Biological Control of Hygrophila: Foreign Exploration for Candidate Natural Enemies

A. Mukherjee1, C. A. Ellison2, J. P. Cuda3, and W. A. Overholt4

1Department of Entomology and Nematology, University of Florida, Gainesville, FL, USA, email: [email protected] 2CABI Europe – UK, Egham, Surrey, United Kingdom, email: [email protected] 3Department of Entomology and Nematology, University of Florida, Gainesville, Florida, USA, email: [email protected] 4Biological Control Research and Containment Laboratory, University of Florida, Fort Pierce, FL, USA, email: [email protected]

Abstract

Hygrophila, Hygrophila polysperma (Roxb.) T. Anders (Acanthaceae) is an invasive aquatic weed of lotic habitats in Florida, USA. This rooted submerged or emergent plant is typically found in flowing fresh water channels and structured shorelines. Hygrophila forms dense vegetative stands that occupy the entire water column, interfering with navigation, irrigation and flood control activities. It is listed as a Federal Noxious Weed and a Florida Exotic Pest Plant Council Category I invasive species. A visible increase in the number of water bodies invaded by hygrophila since 1990 suggested that current methods employed to control this weed are inadequate. A previous study confirmed that hygrophila is a good candidate for classical biological control. However, little information was available on natural enemies affecting hygrophila in its native range. Exploratory field surveys were conducted in a range of habitats in India and Bangladesh during 2008 and 2009. In total, 41 sites were surveyed, including 28 sites in the states of West Bengal and Assam, India and 13 sites in Mymensingh, Bangladesh. The geoposition and altitude of each survey site were recorded. Several collection techniques, e.g. hand picking, Berlese funnel extraction, as well as sweep and clip vegetation sampling, were used to collect natural enemies. A number of insects, including two caterpillars (Precis alamana L., Nymphalidae and an unidentified noctuid moth, Lepidoptera) that defoliate emerged plants, an aquatic caterpillar (Parapoynx bilinealis Snellen, Crambidae, Lepidoptera) feeding an submerged hygrophila, and a leaf mining beetle (Trachys sp., Buprestidae, Coleoptera) were collected during these surveys. In addition, a very damaging aecial rust fungus was collected.

Introduction and a Florida Exotic Pest Plant Council Category I invasive species (FLEPPC, 2009). This aquatic plant escaped cultivation by the aquarium trade Florida historically has been vulnerable to and is now causing serious problems by displacing invasion by exotic animals and plants. The aquatic native aquatic vegetation (Spencer and Bowes, weed hygrophila, Hygrophila polysperma (Roxb.) 1985), primarily in lotic habitats (Spencer and T. Anderson (Acanthaceae), is one such plant. It Bowes, 1985; Sutton, 1995). Hygrophila is believed is listed as a federal noxious weed (USDA, 2006),

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 143 to be native broadly to the southeastern Asiatic from herbaria in the plant’s native range. In India, mainland (Les and Wunderlin, 1981; Spencer and herbaria label information was obtained from The Bowes, 1985; Schmitz, 1990; Angerstein and Lemke, Central National Herbarium, Howrah, India under 1994). This herbaceous perennial weed is capable the Botanical Survey of India. In addition, label of forming dense stands and can occupy the entire information was collected from the Royal Botanic water column, causing disruption in irrigation and Garden Herbarium, Kew, United Kingdom. flood control systems (Schmitz and Nall, 1984; Sutton, 1995). Conventional control techniques do Exploratory Field Surveys in not provide effective control of this invasive weed. Hygrophila’s Native Range Mechanical methods may be useful for removing the floating mats, but harvesting increases the number Exploratory field surveys were conducted in of viable stem fragments that can be transported September 2008 and September 2010 in a range to other areas where they can infest new water of habitats in India and Bangladesh to collect bodies (Sutton, 1995). Established hygrophila natural enemies of hygrophila. In total, 41 sites infestations are also difficult to control with were surveyed, including 28 sites in the states registered aquatic herbicides (Sutton et al., 1994a; of West Bengal and Assam, India and 13 sites in Sutton et al., 1994b). Due to presence of cystoliths Mymensingh, Bangladesh (Fig. 1). The geoposition (calcium carbonate pustules) in its leaves and stems, and altitude of each survey site were recorded. hygrophila is not a preferred food plant for triploid Several collection techniques, e.g., hand-picking, grass carp Ctenopharyngodon idella Val (Cuvier Berlese funnel extraction, sweep and clip vegetation & Valenciennes) (Pisces: Cyprinidae) (Sutton and sampling, as well as dissection of plant parts, Vandiver, 1986). Considering several biological and were used to collect natural enemies. Arthropods economic attributes of this weed, biological control collected during surveys were preserved according may be a viable option for managing hygrophila to standard methods. Specimens were submitted (Pemberton, 1996, Cuda and Sutton, 2000). for identification to cooperating systematists with A noticeable increase in the number of public expertise on specific taxa. lakes and rivers with hygrophila has occurred in Florida since 1990 (Langeland and Burks, 1999). One Results and Discussion of explanations for the recent hygrophila problem is simply a lack of host specific natural enemies attacking the plant, which would give it a competitive Cataloging and Geopositioning Indian Her- advantage over native flora. Surveys of the natural baria Records enemies of hygrophila are needed because there is no information available on potential biological In total, 64 herbaria records of hygrophila were control agents of this aquatic plant (Cuda and examined and the locality information/ ecological Sutton, 2000). The objectives of this study were to notes recorded from the Central National Herbarium (i) catalog and georeference historical populations of at Kolkata, India. The herbarium’s records indicated hygrophila from herbaria records in India; and (ii) that hygrophila was collected from 12 Indian states identify candidate natural enemies associated with and the majority of samples (26 of 64, or 41%) were hygrophila in its native range. from the state of West Bengal in the eastern part of India. The earliest record was dated 1910 and at least one sample was collected at an altitude of 1200 m. Materials and Methods Label information from 41 specimens was collected from the Kew Herbarium, Richmond, Cataloging and Geopositioning Surrey, UK. Records were collected from specimens Herbaria Records dating back to the 1800s and early 1900s from Pakistan, Burma, Vietnam, Taiwan, Sri Lanka In order to identify areas for field surveys, locality and Malaysia. This information was helpful information of hygrophila specimens was collected in delimiting the native range of hygrophila.

XIII International Symposium on Biological Control of Weeds - 2011 144 Session 4 Target and Agent Selection

Exploratory Field Surveys in the Native Range , respectively. Further studies on Trachys sp. will be required in order to understand its life cycle and Insects collected from hygrophila’s native range host range. This information will be essential for belonged to the Orders: Coleoptera (Anthicidae, determining its usefulness as a potential biological Buprestidae, Carabidae, Chrysomelidae, Coccinel- control agent of hygrophila. lidae, Curculionidae, Dytiscidae, Hydrophilidae, A weevil, Bagous luteitarsis Hustache Noteridae, Scarabaeidae, and Staphylinidae), Lepi- (Curculionidae) also was collected during these doptera (Crambidae, Noctuidae and Nymphalidae) surveys (Table 1) (O’Brien and Askevold, 1995). and Hemiptera (Cicadellidae, Delphacidae, Meeno- This is a semi-terrestrial species in the usually plidae and Tingidae). In comparison to its invasive aquatic genus and to date there is no available range, substantially higher numbers of herbivores host plant information (Charles O’Brien, personal were associated with hygrophila in its native range communication). The genus Bagous has been (Mukherjee, 2011). Table 1 lists the insects collected important for classical biological control of aquatic along with their approximate abundances, trophic weeds (O’Brien and Askevold, 1995). B. luteitarsis status or feeding guild and methods of collection. only was collected once during the surveys, suggesting that this is not an abundant species. Coleoptera However, considering that this is a semi-terrestrial species, enabling it to attack both terrestrial and In total, 19 genera of beetles in 11 families were submerged forms of hygrophila, further studies are collected during surveys in India and Bangladesh needed to evaluate its biological control potential. (Table 1). The most important species observed to Several other phytophagous leaf beetles (family: cause direct damage to hygrophila was a leaf mining Chrysomelidae) were collected during this surveys buprestid beetle Trachys sp. (Family: Buprestidae, (Table 1). Congeners of some of these insects have Fig. 2). This insect was collected in both Assam and been considered as potential biological control West Bengal in India, as well as the Mymensingh agents for other weeds (for example, Kok, 2001). area of Bangladesh from terrestrial population However, as indicated in Table 1, these beetles were of hygrophila. Based on field observations, the collected during sweep sampling or Berlese funnel buprestid completes its life cycle within a single leaf extraction. Therefore, there is no direct evidence of hygrophila. The dorsoventrally flattened, wedge- that these insects feed directly and/or exclusively on shaped larva has a characteristic buprestid form hygrophila. with large flattened head and thoracic regions. The larva mines entirely inside hygrophila leaves, feeding Hemiptera on leaf tissue from side to side without damaging the upper and lower leaf epidermises, forming a In total, ten species in the order Hemiptera, transparent leaf cavity. Larval duration is ~3 weeks. including suborders, Auchenorrhyncha (families Pupation occurs within the leaf pocket and the pupal Cicadellidae, Delphacidae and Meenoplidae) and duration is ~7 days. The pupa is brownish in color Heteroptera (family Tingidae) were collected during and ~6-7 mm long and 2-3 mm broad. The adult these surveys (Table 1). All the Auchenorrhyncha beetle is metallic black in color, 3-4 mm in length genera are known to be phloem feeders and are and 1-1.5 mm in width. It is important to note that major pests of agricultural crops including rice the occurrence of this insect was rare. In total, only (Oryza sative L). These insects were collected during 19 larval specimens were collected during the all sweep sampling and probably used hygrophila as the surveys. As this insect completes its life cycle an alternate host. No information on host range within a single leaf, the size of the hygrophila leaf is available on the two tingid species, Belenus may be an important factor limiting its abundance, bengalensis Distant and Cantacader quinquecostatus and damage was usually observed in broader and Fieber, collected during this survey (Distant, longer lower leaves of the plant. The average length 1902;1909). It is noteworthy that several tingids and width of the leaves attacked by this insect ranged have been used as biocontrol agents or reported between 3.0 ± 0.5 cm and 1.2 ± 0.2 cm (mean ± SD) as feeding on invasive weeds in several classical

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 145 weed biological control programs (summarized in 4B). The dorsal surface of the larva changes color Julien and Griffiths, 1998). The above information from green to reddish-brown with successive molts. suggests that tingids could be successful biological Based on field and laboratory observations, this control agents. Further studies will be necessary to insect has five instars and total duration of the larval confirm the feeding damage of B. bengalensis and C. stadium is ~14 days. The larva consumes the entire quinquecostatus on hygrophila. leaf, leaving only the midrib intact (Fig. 4C). Multiple larvae have been observed feeding on the same plant. Lepidoptera Pupation occurs on the plant within a pupal case constructed of 3-4 hygrophila leaves tied together by Three lepidopteran species, an aquatic leaf silk. The pupa is brownish-black in color, ~10 mm cutter Parapoynx bilinealis Snellen (Crambidae: in length and 5 mm in width. Duration of the pupal Acentropinae) (Fig. 3), and two foliage feeders, stage is ~7 days. The adult moth is grayish-brown Nodaria sp. (Noctuidae: Herminiinae) (Fig. 4) and with indistinct blackish antemedial, postmedial and Precis almana L. (Nymphalidae) were collected subterminal wavy lines on the wings; the wingspan during the surveys. All three species were observed ~25 mm. Field observations confirmed that this to cause direct feeding damage to hygrophila. insect can be very damaging, and is probably specific Larvae of P. bilinealis make cases with hygrophila to hygrophila as no feeding was observed on other leaves and feed internally. The mature larva is cream species of nearby plants. Further studies are needed colored, ~15 mm in length and with conspicuous to confirm its host range and determine its biological branched tracheal gills on all body segments except control potential. the prothoracic region (Fig. 3). Presence of branched In addition, a nymphalid butterfly, Precis almana gills along the body segments is the diagnostic L., commonly known as the Peacock Pansy butterfly, character for this genus (Habeck, 1974). Duration was found feeding on hygrophila. This insect, which of the larval stage was ~3 weeks. The larva feeds by was collected from both India and Bangladesh, also scraping leaf tissue from the inner surface of the caused complete defoliation. However, because it case, rendering it transparent. In some cases, they is a polyphagous species known to feed on wide were observed to extend outward from the larval range of plants (Kehimkar, 2008), it has no value for case to feed on nearby leaves. However, no feeding importation biological control of hygrophila. damage to the stem was observed. Pupation occurs within the leaf case, and the leaf case containing the Pathogens pupa floats on the water surface. The duration of the pupal stage was ~7 days. The adult moth (wing span In addition to the aforementioned insects, a ~10 mm, Fig. 3B) is yellowish brown in color with very damaging aecial rust fungus (Pucciniales: Puc- conspicuous white wavy strips on the wings. The ciniaceae, hereafter hygrophila rust) that completely insect caused substantial damage to the submerged killed hygrophila plants was found during surveys in hygrophila plants. Available host records suggest India and Bangladesh (Fig. 5). Rust infections were that insects of this genus tend not to be host specific observed at all locations surveyed, suggesting that it (for example, see Habeck, 1974). Based on available could be an important natural enemy of hygrophila. information, it seems unlikely that P. bilinealis is Historically, rust fungi have been used specific to hygrophila, although our suspicions effectively in classical weed biological control should be confirmed through host range tests. . programs (summarized in Charudattan, 2001). A noctuid moth, Nodaria sp., which attacks both Most rust fungi have a complex life cycle, involving emergent and terrestrial populations of hygrophila, both sexual (aecial and pycnial spores) and asexual was also recorded during these surveys (Fig. 4). This (uredenia, telial and basidial spores) stages (see moth was found at all locations surveyed and was Kolmer et al., 2009 for detailed description of a rust observed to cause complete defoliation of hygrophila life cycle). A prerequisite for a rust fungus to be a plants (Fig. 4C). The fully grown semilooper larva is potential classical biological control agent is that ~35-40 mm long and ~2-3 mm wide with a reddish- it must be autoecious, whereby the rust completes brown dorsal surface and green ventral surface (Fig. its life cycle on a single host species. In case of the

XIII International Symposium on Biological Control of Weeds - 2011 146 Session 4 Target and Agent Selection

hygrophila rust, only the aecial (Fig 5B) and pycnial comm., K.C. Puzari, 2011). Considering the high (Fig. 5C) stages were observed in the field. Repeated level of aeciospore inoculum on H. polysperma, and attempts to inoculate healthy hygrophila plants assuming that the rust is the full cycled P. cacao, with aeciospores from infected plants failed to then it would be expected that uredinia would have initiate infection. However, laboratory observations been found on a native grass, after such long term confirmed that aeciospores germinated on tap-water and intensive observations. Clearly, a more in depth agar medium, producing a normal germ tube (CAE, study of the life cycle is required before this potential personal observation). If this rust were cycling biological control agent is rejected. through aecia alone (microcyclic), then it would Overall, a number of natural enemies, including be expected that the aecia would be behaving as two caterpillars (P. almana and Nodaria sp.) that aeciod-telia, germinating to produce basidiospores. defoliate emerged plants, an aquatic caterpillar (P. However, more work is required to elucidate the bilinealis) feeding on submerged hygrophila, and germination process since this can be influenced by a leaf mining beetle (Trachys sp.) were collected temperature. If this rust is a full-cycled autoecious during surveys in India and Bangladesh, part of the rust, then it would be expected that the aeciospores native range of hygrophila. Some of these insects, would infect the hygrophila, and produce uredinia. in particular, P. bilinealis, Nodaria sp. and Trachys Thus, either the conditions following inoculation sp. hold promise as potential biological control were not conducive to infection, or the rust is indeed agents of hygrophila. Further studies are necessary heteroecious, requiring a primary host to complete to determine their host ranges and specificity to its life cycle. hygrophila. In addition, a very damaging aecial rust In an earlier study, Thirumalachar and fungus (Puccinia sp.) was collected. Although initial Narasimhan (1954) reported that the aecial stage studies suggested that this rust could be the aecial of the heteroecious leaf rust, Puccinia cacao McAlp. stage of the heteroecious P. cacao, detailed cross- occurs on Hygrophila spinosa (Acanthaceae), a inoculation studies, involving its primary host H. congener of H. polysperma. The common grass compressa, are necessary to confirm its identity. Hemarthria compressa (L.f.) R.Br. (Poaceae) is the primary host of P. cacao. They reported that on H. Acknowledgements spinosa, the rust develops a systemic infection and the infected shoots are paler in color. Interestingly, This research was supported by grants from similar observations were made on hygrophila. the Environmental Protection Agency (Grant Laundon (1963) provided descriptions of all rust ID: X7-96433105) as part of the Osceola County fungi infecting Hygrophila spp., at that time. When Demonstration Project on Hydrilla and Hygrophila compared, the measurements of aecial and pycnial in the Upper Kissimmee Chain of Lakes, FL, and the spores, collected from H. phlomoides Nees, H. Florida Fish and Wildlife Conservation Commission, salicifolia (Vahl) Nees and H. Spinosa T. Anders as Invasive Plant Management Section (Contract No. reported in Laundon (1963), match closely with SL849 –UFTA120). Our thanks go to Dr. K.C. Puzari those of the hygrophila rust that we collected (Table for his support in the undertaking of the field work 2). These results also suggest the rust from H. in Assam and his detailed field observations. polysperma could be the sexual stage of a different pathotype of P. cacao. Cross-inoculation studies to test the ability of aeciospores from H. polysperma to References initiate infection (uredinia) in a currently unknown primary host (possibly a grass such as Hemarthria Angerstein, M.B. & Lemke, D.E. (1994) First records compressa) is necessary to confirm the identification of the aquatic weed Hygrophila polysperma of this rust fungus. However, field observations in (Acanthaceae) from Texas. Sida Contributions to Assam, over two growing seasons and at multiple Botany 16, 365–371 natural sites, have not resulted in the discovery Charudattan, R. (2001) Biological control of weeds of a grass infected with uredinia, growing in the by means of plant pathogens: Significance for vicinity of aecia-infected H. polysperma (pers. integrated weed management in modern agro-

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 147

ecology. Biocontrol 46, 229–260 Mukherjee A. (2011) Prospects for classical Cuda, J.P. & Sutton, D.L. (2000) Is the aquatic biological control of the aquatic invasive weed weed hygrophila, Hygrophila polysperma Hygrophila polysperma (Acanthaceae). PhD (Polemoniales: Acanthaceae), a suitable target for Disseration, University of Florida. classical biological control? Proceedings of the X O’Brien, C. & Askevold, I. (1995) Systematics International Symposium on Biological Control and evolution of weevils of the genus Bagous of Weeds, Bozeman, Montana, USA, 4-14 July, germar (Coleoptera: Curculionidae). V. 1999, 337–348 Taxonomic treatment of the species of the Indian Distant, W. (1902) Rhynchotal notes. XIII. subcontinent. Contributions of the American Heteroptera: Families Tingididae, Phymatidae Entomological Institute (USA) 28, 1–184 and Aradidae. Annals And Magazine of Natural Pemberton, R.W. 1996. The potential of biological History 9, 353–362 control for the suppression of invasive weeds of Distant, W. (1909) New Oriental Tingididae. Annales southern environments. Castanea 61, 313–319. de la Socié té Entomologique de Belgique 53, Schmitz, D.C. & Nall, L.E. (1984) Status of Hygrophila 113–123 polysperma in Florida. Aquatics 6, 11–12–14 FLEPPC (2009) List of Invasive Plant Species. p 4, Schmitz, D.C. (1990) The invasion of exotic aquatic Florida Exotic Pest Plant Council and wetland plants in Florida: History and efforts Habeck, D.H. (1974) Caterpillars of Parapoynx in to prevent new introductions. Aquatics 12, 6–7– relation to aquatic plants in Florida. Hyacinth 16 Control J 12, 15–8 Spencer, W. & Bowes, G. (1985) Limnophila Julien, M.H. & Griffiths, M.W. (1998) Biological and hygrophila - a review and physiological control of weeds, Fourth edition: A World assessment of their weed potential in Florida. Catalogue of Agents and their Target Weeds. Journal of Aquatic Plant Management 23, 7–16 CABI Publishing, Wallingford, Oxfordshire, UK Sutton, D.L. & Vandiver, V.V.J. (1986) Grass carp: x+223p pp A fish for biological management of hydrilla Kehimkar, I. (2008) The Book of Indian Butterflies. and other aquatic weeds in Florida. Florida Bombay Natural History Society. Oxford Agriculture Experiment Station Bulletin 867 University Press, UK. Sutton, D.L. (1995) Hygrophila is replacing hydrilla Kok, L. (2001) Classical biological control of nodding in South Florida. Aquatics 17, 4–6, 8, 10 and plumeless thistles. Biological Control 21, Sutton, D.L., Bitting, L.E. & Moore, W.H. (1994a) 206–213 Winter treatment with endothall for control of Kolmer, J.A., Ordonez, M.E. & Groth, J.V. (2009) East Indian hygrophila in South Florida canals. The Rust Fungi. In: Encyclopedia of Life Sciences Aquatics 16, 4–6, 8 (ELS). John Wiley & Sons, Ltd. Chichester. Sutton, D.L., Bitting, L.E., Moore, W.H. & Baker, G.E. Langeland, K.A. & Burks, K.C. (1999) Identification (1994b) Summer treatment of Hygrophila with and biology of non-native plants in Florida’s endothall in South Florida. Aquatics 16, 4–6, 8 natural areas. University of Florida, Gainesville, Thirumalachar, M. & Narasimhan, M. (1954) Florida Morphology of spore forms and heteroecism in Laundon, G.F. (1963) Rust fungi, I. On Acanthaceae. Puccinia cacao. Mycologia 46, 222–228 Mycological papers, C.M.I. 89, 1–89 USDA (2006) Federal Noxious Weed List. United Les, D.H. & Wunderlin, R.P. (1981) Hygrophila States Department of Agriculture. http://plants. polysperma Acanthaceae in Florida, USA. Florida usda.gov/java/noxious?rptType=State&statefi Scientist 44, 189–192 ps=12

XIII International Symposium on Biological Control of Weeds - 2011 148 Session 4 Target and Agent Selection

Table 1. Insects collected from hygrophila during surveys in India and Bangladesh, native habitats of Hygrophila polysperma. Approx. Trophic level or Methods of Taxonomy abundance1 feeding guild2 collection3 Anthicidae Anthelephila mutillaria Saunders R Scavenger? Sweep Buprestidae Trachys sp. U Leaf miner Hand coll. Carabidae Bembidion sp. R Predator Sweep Clivina sp. Latreille U Predator Sweep Tachys sp. R Predator Sweep Chrysomelidae Altica sp. U Leaf feeder Sweep Aspidomorpha sp. R Leaf feeder Sweep Cassida sp. A U Leaf feeder Sweep Cassida sp. B U Leaf / root feeder Sweep Chaetocnema sp. R Leaf feeder Sweep Lema sp. A R Leaf feeder Sweep Lema sp. B R Leaf feeder Berlese Pachnephorus sp. U Leaf feeder Sweep Philipona sp. U Leaf feeder Sweep Coccinellidae Harmonia sp. C Predator Sweep Curculionidae Bagous luteitarsis Hustache R ? Berlese Hydrophilidae Helochares sp. U Predator Berlese Paracymus sp. U Predator Berlese Regimbartia sp. U Predator Berlese Scarabaeidae Rhyssemus sp. Mulsant R Scavenger? Sweep Staphylinidae Philonthus sp. Stephens R Predator Sweep Hemiptera Cicadellidae Nephotettix sp. C Sap feeder Sweep Cofana spectra Distant C Sap feeder Sweep Cofana unimaculata Signoret C Sap feeder Sweep

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 149

Table 1. Insects collected from hygrophila during surveys in India and Bangladesh, native habitats of Hygrophila polysperma. Approx. Trophic level or Methods of Taxonomy abundance1 feeding guild2 collection3 Hecalus sp. U Sap feeder Sweep Scaphomonus indicus Distant U Sap feeder Sweep Delphacidae Nilaparvata sp. C Sap feeder Sweep Perkinsiella sp. Kirkaldy C Sap feeder Sweep Tingidae Belenus bengalensis Distant R Sap feeder Sweep Cantacader quinquecostatus Fieber R Sap feeder Sweep Lepidoptera Crambidae Parapoynx bilinealis Snellen R Leaf cutter Hand coll. Noctuidae Nodaria sp. C Leaf feeder Hand coll. Nymphalidae Precis almana L. C Leaf feeder Hand coll. 1Approximate abundance: R = Rarely collected, 3 times or less; U = Uncommonly collected, 4-10 times; and C = Commonly collected, >10 times 2Trophic level and feeding guild information for herbivorous species does not imply that species included on this list were actually observed using hygrophila as a food source 3Methods of collection: Sweep = Sweep net sampling; Berlese = Berlese funnel extraction and Hand coll. = larva collected from field and reared to adult

XIII International Symposium on Biological Control of Weeds - 2011 150 Session 4 Target and Agent Selection

Table 2. Comparison of aecial rust collected from Hygrophila polysperma with that collected from H. phlomoides, H. salicifolia and H. spinosa by Laundon, (1963) Aecia (diam.) Aeciospores (diam.) Pycnia (diam) Pycniospores Puccinia sp. collected from 228.1 - 398.1 µm 18.43 – 28.4 µm 103.4 - 120.1 µm Paraphysate hygrophila

Laundon (1963) 200 - 400 µm 15 - 30 µm 80 - 100 µm Paraphysate

Figure 1. Survey sites in India (n = 28) and Bangladesh (n = 13). In India, surveys were conducted in the states of West Bengal (n = 15) and Assam (n = 13).

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 151

Figure 2. Trachys sp. (Coleoptera: Buprestidae) collected from H. polysperma; Larva mining inside the leaf cavity, inset: close-up of larva (A); Pupa inside the leaf cavity (B); Ventral and dorsal views of the adult beetle (C-D). Photo credit A. Mukherjee

Figure 3. Larva (A) and adult (B) of Parapoynx bilinealis (Lepidoptera: Crambidae) collected from H. polysperma. Note the presence of branched tracheal gill on larval body – a characteristic of the genus. Photo credit A. Mukherjee

XIII International Symposium on Biological Control of Weeds - 2011 152 Session 4 Target and Agent Selection

A B

C D

Figure 4. Nodaria sp. (Lepidoptera: Noctuidae) collected from H. polysperma; Larva feeding on hygrophila leaves (A); Pupa (B); Feeding damage (C); Adult moth (D). Photo credit A. Mukherjee A B

Figure 5. Rust fungus (Puccinia sp.) collected from H. polysperma. Rust infected hygrophila plant (A); Cross section of aecia (B); Cross section of pycnia (C). Photo credit A. Mukherjee

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 153

Biological Control of Rubus alceifolius (Rosaceae) in La Réunion Island (Indian Ocean): From Investigations on the Plant to the Release of the Biological Control Agent Cibdela janthina (Argidae)

T. Le Bourgeois1, S. Baret2 and R. D. de Chenon3

1Cirad, UMR AMAP, TA A51/PS2, Boulevard de la Lironde, F34398 Montpellier Cedex 5, France. [email protected] 2Parc National de la Réunion, 112 rue Sainte-Marie, 97400 Saint-Denis, La Réunion, France. ste- [email protected] 3Pusat Penelitian Kelapa Sawit, Balai Penelitian Marihat, P.O.Box 37, Sumatera Utara, Indonesia. [email protected]

Abstract

The giant bramble (Rubus alceifolius Poir.: Rosaceae), native to Southeast Asia, is one of the most invasive plants in La Réunion. A ten year research program was launched in 1997 with three components: i) genetic diversity, ii) development strategy, and iii) selection of biological control agents. Introduced populations in La Réunion, Mauritius, Mayotte and Australia were clonal and far from the highly variable native populations in Asia, while Madagascar populations appeared intermediate. Seed production is by apomixis in La Réunion Island and by allogamy in the native habitat. Fruit production occurs up to 1,100 m elevation while vegetative multiplication is possible up to 1,700 m. The plant grows in well lighted places, invading forest edges, and all open areas. From surveys of Rubus natural enemies in its native range, the sawfly Cibdela janthina (Klug) (Argidae) was selected as the most promising biological control agent and studied. The first population was thus released in La Réunion in early 2008 with the agreement of the local authorities for the biological control of R. alceifolius. It is now naturalized, spreading and under evaluation.

Introduction transformation with the introduction of more than 2000 plant species, of which 628 have become naturalised (Lavergne et al., 1999). During their Giant bramble (Rubus alceifolius Poir.: Rosaceae) evaluation of the threat posed by invasive plants to is an invasive Southeast Asian bramble introduced th the island, Macdonald et al. (1991) underlined that to La Réunion Island (Indian Ocean) in the mid 19 62 exotic species were major threats, among them century. In 1892, it was already cited as fatal for the was R. alceifolius, which topped the list. island (Lavergne, 1978), and Rivals (1960) said “it For many years, the Office National des Forêts was a real pain for natural environment. Density was attempted to bring the weed under control. But so high that regeneration of indigenous forest plants neither mechanical weeding nor chemical control was impossible under Rubus thickets.” At that time, he was successful. Control was possible only on small also mentioned that “only a biological solution could surface areas, and needed to be repeated regularly. destroy this plant and solve the problem”. Also, the cost was extremely high (at present, €2 Since the 16th century, the native ecosystems million is spent annually on the control of invasive of La Réunion Island have undergone rapid plants) and the use of herbicides in native forests

XIII International Symposium on Biological Control of Weeds - 2011 154 Session 4 Target and Agent Selection

has raised ecotoxicological concerns. In 1992, the Materials and Methods Conseil Régional of La Réunion made R. alceifolius a priority and decided to fund a research program La Réunion Island (2,512 km² in area) is a to develop integrated control methods, with the volcanic island in the Indian Ocean rising to 3,061 m emphasis placed on biological control. This program a.s.l., with considerable climate variations. Rainfall started in 1997 and ended in 2006 with the selection ranges from 8,000 mm per year on the east coast to of a biological control agent. It was then followed 500 mm per year on the lowland west coast (Robert, by complementary action funded by the DIREN 1986). The island’s vegetation comprises four main (Direction Régionale de l’Environnement) of La types that are determined by a combination of Réunion until 2009 for the introduction, acclimation rainfall and temperature: semi xerophytic tropical and release of the selected biological control agent forest, lowland humid tropical forest, mountain Cibdela janthina (Klug) (Hymenoptera: Argidae). humid tropical forest and ericoid vegetation at high While much research has been done on the elevations (Cadet, 1977). biological control of other species of Rubus (Bruzzese The giant bramble, Rubus alceifolius Poir. and Lane, 1996; Evans et al., 1999; Gardner et al., (Rosaceae, subgenus Malachobatus) was first 1997; Groves et al., 1997; Julien and Griffiths, 1998; described as a subspecies of R. moluccanus L., but Nagata and Markin, 1986), little work has been done is now considered a separate species (Kalkmann, on tropical Rubus species or R. alceifolius itself. It 1993). It is a shrub with arching or climbing concerned its distribution, seed production and branches up to 5 m long in native areas or up to 15 spread in La Réunion (Sigala and Lavergne, 1996; m in length in areas of introduction. Branches and Strasberg, 1995; Thebaud, 1989). The chrysomelid leaves bear curved prickles and yellow hairs. Stipules Phaedon fulvescens Weise, originating in Northern are large, orbicular, and deeply digitately divided. Vietnam, was mentioned as a potential candidate Leaves are simple, orbicular to broadly ovate, 10-30 for biological control (Jolivet, 1984). Other natural cm in diameter, with 5-7 lobes, and cordate at the enemies were an unidentified stick insect and two base. Inflorescence is terminal, consisting of up to 4 rust fungi, Hamaspora acutissima P.Syd. & Syd. racemes with up to 8 flowers. Flower bud is globular, and Gerwasia rubi Racib. recorded in Thailand by petals are orbicular, white. Collective fruit is globular, Taysum et al. in 1991 (unpublished document). All succulent, 1 cm in diameter and containing many previous experiments carried out worldwide in the red drupelets (Kalkmann, 1993). The giant bramble biological control of Rubus spp., used pathogens. is exotic in Australia (Queensland), Mauritius, The project was built around three components. Mayotte, Madagascar and La Réunion Island but is The first concerned the plant’s genetic diversity: native to Southeast Asia (southeast China, Hainan, i) to compare the diversity between the area of Taiwan, Myanmar, Thailand, Laos, Cambodia, introduction and the native range in order to match Vietnam, Indonesia/Sumatra, Java, Malaysia, Lesser genetic and geographic origins of the invader, and Sunda Islands, Borneo and Sulawesi) (Friedmann, 1997; Kalkmann, 1993; Parsons and Cuthbertson, ii) to determine the degree of specificity needed by 1992). At La Réunion it frequently occurs as large biological control candidates and the opportunity for stands along forest edges, road and river sides from their use throughout the area of introduction. The sea level up to 1,700 m of elevation (Baret et al., second component was the biological study of the 2004). weed under La Réunion environmental conditions in Genetic diversity studies of R. alceifolius were order to highlight factors for growth, multiplication based on 224 specimens: 116 from introduced and spread. The third component aimed to select populations (La Réunion (75), Mayotte (8), Mauritius and study the biological control candidates suitable (7), Madagascar (19) and Australia/Queensland (7), for introduction and release in La Réunion. and 108 from native populations (Thailand (59), This paper presents a compilation of the project’s Vietnam (30), Laos (1), Java (4), and Sumatra (14). main findings leading to the proposal for a biological Thirty specimens of other Rubus species were also control of the giant bramble in La Réunion using the evaluated (La Réunion (3), Thailand (12), Vietnam sawfly Cibdela janthina. (10), Laos (2), and Sumatra (3). Plant DNA was

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 155 extracted and processed following the protocol most promising ones were collected alive and reared of Bousquet et al. (1990). Genetic differentiation in laboratories in Sumatra, Montpellier and La was assessed by amplified fragment length Réunion for further biological and specificity studies polymorphism (AFLP), using 4 primer pairs and in order to select those most suitable for biological the restriction enzymes EcoRI and MseI, as detailled control of R. alceifolius at La Réunion. by Amsellem et al. (2000). The genetic distance between individuals was calculated and expressed as the Simple Matching distance (Rohlf, 1993). A tree Results was constructed according to the Neighbour-Joining Method (Felsenstein, 1993). The study focused on Rubus genetic diversity and putative origins two levels of comparison: individuals and areas of R. alceifolius (Amsellem et al., 2000). The reproductive biology of R. alceifolius in the The genetic study of R. alceifolius revealed two native and introduced habitats was assessed and well-separated groups, both distinct from the compared using microsatellite markers specific for out-group corresponding to other Rubus species. Rubus species (Amsellem et al., 2001a). We compared The first group consisted of all the samples taken the reproductive system (fruit set) of R. alceifolius within the native range and the second group, in its native range (nine half-sib seedlings derived all the samples from the area of introduction. A from different fruits on a single plant sourced from study of the within-area diversity showed relatively Vietnam), in its area of introduction in Madagascar marked genetic diversity between individuals from (three individuals and their half-sib progeny from countries of the native range. On contrary, each two localities), and in La Réunion where 44 flowers population sampled in the Indian Ocean islands (La from several parents were manually fertilized with Réunion, Mayotte, Mauritius, and Australia), with pollen from other plants (Amsellem et al., 2001b). the exception of Madagascar, was characterized by Developmental patterns were then studied a single different R. alceifolius genotype. All these at La Réunion by conducting architectural and genotypes were closely related to individuals from morphometric analyses that described individuals Madagascar where polymorphism was intermediate at the five specific growth stages from seedlings between the situation in areas of introduction and to mature plants (Baret et al., 2003a; Baret et al., the native range. 2003b). Altitudinal variations in fertility and Many authors have discussed the different vegetative growth were assessed by counting flowers, hypothetical origins of this exotic plant. E. Jacob de fruits, seeds and leaves in eight randomly located Cordemoy (1895) mentioned that it may have been quadrats at six sites from 50 m to 1,500 m a.s.l., and introduced at La Réunion in 1846. Nevertheless a by estimating the soil seed bank (Baret et al., 2004). previous text “Ralliement du 17 août 1892” cited by Biological control agents were selected from Lavergne (1978) specified that this plant came from surveys carried out in the native range (Vietnam, the Botanical Garden in Calcutta. Other authors Laos, Thailand, China, Indonesia (Sumatra) and thought it was sent from Vietnam (Jolivet, 1984), Singapore) from 1997 to 2004, and in the area or that it may either have been introduced first in of introduction at La Réunion Island (1998). Mauritius by Commerson on his return from Java Altogether, stands of Rubus were examined at 309 in 1768 (Rivals, 1960), or introduced into Mauritius different locations subject to different climatic from Madagascar (Vaughan, 1937) and then spread conditions, ranging from the lowlands to the highest to La Réunion in the 1850s (Rivals, 1960). Our elevations (2,500 m a.s.l. in Doi Inthanon Mountains, results suggest that R. alceifolius was first introduced Northern Thailand) and from equatorial (Toba Lake, into Madagascar, perhaps on multiple occasions. The Northern Sumatra) to tropical climates with fairly Madagascan individuals were thus the immediate cold winters (Guangdong and Hainan mountains source of the plants that colonized other areas of in China). Rubus natural enemies (arthropods or introduction. Successive nested founder events pathogens) were collected and identified. Symptoms appear to have resulted in the populations in the and environmental conditions were recorded. The area of introduction showing a cumulative reduction

XIII International Symposium on Biological Control of Weeds - 2011 156 Session 4 Target and Agent Selection

in genetic diversity. The marked genetic differences development, whereas it “explores” the environment between the populations in the native range and during the adult stage (Baret et al., 2003b). in the area of introduction prevented us from To determine the invasive capacity of R. determining the geographical origin of the alien alceifolius, fertility and vegetative growth were plant (Amsellem et al., 2000). studied at different altitudes on La Réunion Island (Baret et al., 2005). Flowering period duration, seed Genetic diversity and the weed’s biological production, and the seed bank were found to be strategy as pointers for control negatively correlated with elevation (50 – 1,500 m a.s.l.). At a lowland site, fruit production averaged The low diversity of the populations found in most between 30 and 80 fuits m-², while no fruits were areas of introduction suggested that a biological observed above 1,100 m. The seed bank was greater control agent efficient against one individual should under R. alceifolius patches (>10,000 seeds.m-²) be able to attack the entire population on the island. than in understories not colonized by the bramble Both the genetic diversity patterns and (approximately 3,000 seeds.m-²). Seed dispersion in differences between half-sib progeny and their forest was mainly by running water. Although the maternal parents (revealed by microsatellite number of leaves per unit area was similar along markers) showed that in the native range, seeds the entire gradient studied, the reduced fruiting are produced sexually (Amsellem et al., 2001a; in upland areas might be offset by an increase in Amsellem et al., 2001b. By contrast, in Madagascar, vegetative growth. Monospecific bramble patches in over 85% of the half-sib progeny resulting from lowland areas may serve as the sources of seed for open pollination gave multilocus genotypes the colonization of new areas by bird dissemination. identical to those of their respective maternal Once established at high elevations, the weed grows parents. Seeds thus appear to be produced mostly vegetatively without flowering and multiply by or exclusively by apomixis in Madagascar. We layering, cutting or sucker (Baret et al., 2004). therefore suggest that Madagascan populations resulted from the hybridization of an introduced Selection of potential biological control R. alceifolius and native populations of presumably agents R. roridus Lindley that Kalkman (1993) considered similar and synonymous to R. alceifolius. Apomixis Fifty one arthropods and four pathogens were was therefore a consequence of this hybridization. recorded and collected during the surveys conducted In Reunionese populations of R. alceifolius, seeds in the native range and in La Réunion Island. obtained in controlled pollination experiments were Particular care was taken to select agents on the all genetically identical to the maternal parents. basis of a combination of criteria (type of damage While genetic variation (microsatellite markers) in to and impact on R. alceifolius, host specificity, life Reunionese populations was low, it was sufficient traits etc.) for further biological and specificity to demonstrate that seeds could not have resulted studies. Of the leaf feeders, sawflies from Sumatra from fertilization by the pollen donors chosen for and China (Cibdela janthina, C. chinensis Rohwer, controlled pollinations, or from autogamy, but were and Arge siluncula Konow) appeared to be the produced exclusively by apomixis (Amsellem et al., most promising. They caused complete defoliation 2001b). This phenomenon was responsible for the of the weed and seemed to be highly specific. The clonal population in the area of introduction. beetles Phaedon fulvescens and Cleorina modiglianii The architectural and morphological studies Jacoby, found respectively in Vietnam and Sumatra, showed five developmental stages for R. alceifolius, were also promising. The rust fungus Hamaspora differing by several markers such as internode acutissima was observed and collected in many length and diameter, pith diameter, and plant shape. places throughout Asia. Only a few common insects A heteroblastic developmental pattern was thus were reported in La Réunion Island, but never revealed for the plant, midway between that of a damaging R. alceifolius. bush and a vine. The results also showed that this Phaedon fulvescens Weise (Coleoptera: species taps environmental resources early in its Chrysomelidae) was collected in Northern Vietnam

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 157 while feeding on Rubus spp. leaves on the edge of The full life cycle ranged from 48 to 62 days. The mountain forests at 900 m a.s.l. Field observations insect may complete six generations per year coupled with biological and host specificity studies in without any diapause. Adult life span was only 7-14 Montpellier generated a wealth of new information days and they were found not to be feeding under about this species. Although not encountered Sumatran environmental conditions; only drinking frequently, individuals were very numerous in the dewdrops on leaves, while the larvae were feeding on populations observed and both adults and larvae R. alceifolius leaves consumed systematically along damaged the plant by leaf skeletonizing. Eggs were the branches in a top-down process. Host range laid separately on the underside of leaves and coated tests were conducted on 41 plant species from 13 with feces. We also noted that the insect undergoes a botanical families chosen on the basis of phylogeny two-month summer diapause. Therefore, this insect and economic or conservation issues for La Réunion. may only complete a single generation in a year, not Cibdela janthina appeared to have a very narrow host three or four as initially thought. Field observations range, feeding only on Rubus species. Starvation indicated that the beetle was highly specific to Rubus tests showed some feeding on certain subspecies species of the Malachobatus subgenus. But host of R. moluccanus not present in La Réunion, on R. specificity tests carried out on R. apetalus (indigenous fraxinifolius (exotic to La Réunion) and on R. apetalus of La Réunion) of the Ideobatus subgenus showed (indigenous in La Réunion). Choice tests showed that the insect can also feed on this plant and survive that the insect mainly prefer to feed on R. alceifolius. throughout its lifecycle. We therefore rejected P. Temperature conditions that impact on the insect’s fulvescens for the biological control of R. alceifolius development should keep C. janthina under 1,000 m in La Réunion (Le Bourgeois et al., 2004). of elevation while R. apetalus is present from 700 m Cleorina modiglianii Jacoby (Coleoptera: to 1,700 m. Considering these results and the insect’s Chrysomelidae) was found in Sumatra on several biological features, C. janthina was considered as a Rubus spp. in the shade provided by Pinus merkusii good potential biological control agent against R. Jungh. & de Vriese forests from 700 to 1,200 m. alceifolius in La Réunion. Accordingly, a petition was This beetle caused leaf skeletonizing damage to R. made for permission to introduce and release it. alceifolius and R. moluccanus L. We were unable to Hamaspora acutissima P.Syd. & Syd. (Uredinales: find eggs or larvae despite numerous field surveys. Phragmidiaceae) was observed at many locations in Host specificity tests were carried on adults in the Vietnam, Thailand and Indonesia (Sumatra). It was laboratory, comparing R. alceifolius (Réunion), R. alceifolius (Sumatra), R. apetalus (L Réunion) and R. visible on the upper face of leaves as small brown to fraxinifolius (La Réunion). The adults only fed and yellowish spots and on the lower face as bunches of survived on R. alceifolius (Sumatra) and R. apetalus orange paraphyses containing teliospores. Spots were (La Réunion). This insect was therefore rejected as a found on isolated leaves and plants, or as intense potential candidate for biological control. infestations covering all parts of Rubus plants. In Cibdela janthina (Klug) (Hymenoptera: Argidae) cases of marked contamination, leaves were drying was recorded in Sumatra. The insect’s behavior and withering. All the Rubus mentioned as infected was observed in the field while its biological traits by this fungus belonged to the Malachobatus were assessed in the laboratory in Sumatra. Mating subgenus indicating the narrow host range of the rust. happened in full light at 30°C and 80% humidity Biological and host specificity studies showed that two days after female emergence. Then eggs were Asian Rubus species belonging to the Malachobatus inserted into the main nerves of the plant’s upper subgenus could sometimes be inoculated. R. young leaves not yet fully opened. Average fertility alceifolius from La Réunion was inoculated but the was 58 eggs per female, with 84% viability. The pathogen stopped growing at the mycelium stage eggs hatched after 10 days of incubation. The larvae without producing new teliospores. Rust fungi are completed seven instars within 25 to 30 days, and known to be highly specific (Evans and Gomez, then pupated in a silk cocoon under the leaf litter. 2004). Our results confirmed that the Reunionese Larvae were gregarious during the major part of R. alceifolius is genetically too different from those their development and presented a typical S-shape. in the native range to allow H. acutissima attack.

XIII International Symposium on Biological Control of Weeds - 2011 158 Session 4 Target and Agent Selection

Conclusion Agricultural Science Institute for their help during the surveys conducted in Vietnam, and the staff of the Guangdong Entomological Institute in China The alien invasive giant bramble found in La for their help during the surveys conducted in Réunion and other areas of introduction is genetically China and in Hainan Island. They are very thankful different from those in the native range and probably to David Smith from the Smithonian Institute for resulted from hybridization in Madagascar. It has identification of the sawflies; Jean François Voisin a high growth potential and produces apomyctic from the Museum National d’Histoire Naturelle fruits in the area of introduction. With its marked for identification of weevils; Serge Quilici from phenotypic plasticity, the plant is able to fruit below Cirad for Reunionese insect identification; Haruo 1,100 m a.s.l. and can grow and multiply vegetatively Matsuzawa from Hasuda, Japan for C. modiglianii up to 1,700 m. identification, Pierre Jolivet for identification of Of the biological agents collected in the bramble’s the chrysomelid beetles and Laurence Dedieu from native range, the sawfly C. janthina showed the best Cirad for reviewing this paper. biological and ecological traits and host specificity that could justify its introduction into La Réunion to regulate R. alceifolius populations under 1000 m References a.s.l. It also appeared to have the most severe impact on weed growth. Amsellem, L., Noyer, J.L., Le Bourgeois, T. & Therefore, a petition form to introduction and Hossaert-Mckey, M. (2000). Comparison of release of C. janthina was submited in 2006 to the genetic diversity of the invasive weed Rubus ad hoc scientific committee and was accepted by alceifolius Poir. (Rosaceae) in its native range local authorities. Cocoons from Indonesia/Sumatra and in areas of introduction, using amplified were introduced in mid 2007 for the rearing of the fragment length polymorphism (AFLP) markers. sawfly and acclimatization at La Réunion. The first Molecular Ecology 9, 443–455. population was released on the east coast of La Amsellem, L., Dutech, C. & Billote, N. (2001a). Réunion in early 2008. It is now spreading well, and Isolation and characterisation of polymorphic controlling populations of the giant bramble. Studies microsatellite loci in Rubus alceifolius Poir. of the spread dynamics and impact of C. janthina on (Rosaceae), an invasive weed in La Réunion R. alceifolius thickets at the island level are ongoing, as Island. Molecular Ecology Notes 1, 33–35. is a study of the impact of the decline of populations Amsellem, L., Noyer, J.L. & Hossaert-Mckey, M. of the giant bramble on natural vegetation dynamics. (2001b). Evidence for a switch in the reproductive biology of Rubus alceifolius (Rosaceae) towards Acknowledgements apomixis, between its native range and its area of introduction. American Journal of Botany 88, 2243–2251. This research program was funded by the Baret, S., Nicolini, E., Humeau, L. & Le Bourgeois, T. Conseil Régional of La Réunion (REG/97/0307 (2003a). Use of architectural and morphometric and REG/2004 0106) and then by the DIREN. analysis to predict the flowering pattern of the The authors wish to thank partner institutions invasive Rubus on Réunion island (Indian Ocean). such as the Centre de Coopération Internationale Canadian Journal of Botany 81, 1293–1301. en Recherche Agronomique pour le Développement Baret, S., Nicolini, E., Le Bourgeois, T. & Strasberg, (Cirad) (France), the Université de La Réunion D. (2003b). Development patterns of the invasive (France), the National Biological Control Research bramble (Rubus alceifolius Poiret: Rosaceae) Center of Kasetsart University (Thailand), the in Réunion Island: An architectural and Centre National de la Recherche Scientifique (France), morphometric analysis. Annals of Botany 91, and the Indonesian Oil Palm Research Institute 39–48. (Indonesia). They also wish to thank the National Baret, S., Maurice, S., Le Bourgeois, T. & Strasberg, Institute of Plant Protection and the Vietnam D. (2004). Altitudinal variation in fertility and

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 159

vegetative growth in the invasive plant Rubus Jolivet, P. (1984). Phaedon fulvescens Weise (Col. alceifolius Poiret (Rosaceae), on Réunion island. Chrysomelidae Chrysomelinae) un auxilliaire Plant Ecology 172, 265–273. possible dans le controle des Rubus aux tropiques. Baret, S., Radjassegarane, S., Le Bourgeois, T. & Bulletin de la Société Linéenne de Lyon 53, 235– Strasberg, D. (2005). Does the growth rate of 246. different reproductive modes of an introduced Julien, M.H. & Griffiths, M.W. (1998). Biological plant cause invasiveness? International Journal of Control of Weeds. A world catalogue of agents and Botany 1, 5–11. their target weeds. CABI Publishing, Wallingford, Bousquet, J., Simon, L. & Lalonde, M. (1990). DNA UK. 223 p. amplification from vegetative and sexual tissues of Kalkmann, C. (1993). Rosaceae, Flora Malesiana, trees using polymerase chain reaction. Canadian Series I, Spermatophyta: Flowering Plants, Journal of Forestry Research 20, 254–257. Foundation Flora Malesiana, Rijksherbarium, Bruzzese, E. & Lane, M. (1996). The Blackberry Leiden, The Netherlands, pp. 227–351. Management Handbook. KTRI, Melbourne, Lavergne, C., Rameau, J.C. & Figier, J. (1999). The Australia. 50 p. invasive woody weed Ligustrum robustum subsp. Cadet, T. (1977). La végétation de l’île de La Réunion: walkeri threatens native forest on La Réunion. Etude Phytoécologique et Phytosociologique. Biological Invasions 1, 377–392. Thèse de doctorat, Université d’Aix-Marseille, Lavergne, R. (1978). Les pestes végétales de l’île de la Marseille, France. 362 p. Réunion. Info-Nature 16, 9–58. Evans, K.J., Symon, D.E., Hosking, J.R., Mahr, Le Bourgeois, T., Goillot, A. & Carrara, A. (2004). New F.A., Jones, M.K. & Roush, R.T. (1999) Towards data on the biology of Phaedon fulvescens (Col. improved biocontrol of blackberries, In Chrysomelinae), a potential biological control Proceedings of the. 12th Australian Weeds agent of Rubus alceifolius (Rosaceae). In New Conference, (eds. Bishop, A.C., Boersma, M. & contributions to the biology of Chrysomelidae Barnes, C.D.), pp. 325–329. Hobart, Tasmania, (eds. Jolivet, P.H., Santiago-Blay, J.A. & Schmitt, Australia M.), pp. 757–766. SPB Academic Publishers, The Evans, K.J. & Gomez, D.R. (2004). Genetic markers Hague, Netherlands. in rust fungi and their application to weed Macdonald, I.A.W., Thébaud, C., Strahm, biocontrol. In Genetics, Evolution and Biological W.A. & Strasberg, D. (1991). Effects of alien Control (eds. Ehler, L.E., Sforza, R. & Mateille, invasions on native vegetation remnants on La T.), pp. 73–96. CABI publishing, Trowbridge, UK. Réunion (Mascarenes Islands, Indian Ocean). Felsenstein, J. (1993). PHYLIP (Phylogeny Inference Environmental Conservation 18, 51–61. Package) version 3.5c. Distributed by the Nagata, R.F. & Markin, G.P. (1986) Status of insects author. Department of Genetics, University of introduced into Hawaii for the biological control Washington, Seattle, USA. of the wild blackberry Rubus argutus Link., In Friedmann, F. (1997). Flore des Mascareignes - La Proceedings of the 6th Conference in Natural Réunion, Maurice, Rodrigues - 81. Rosaceae. Sciences, pp. 53–64. Manoa, Hawaii ORSTOM, Paris, France, pp. 1–11. Parsons, W.T. & Cuthbertson, E.G. (1992). Noxious Gardner, D.E., Hodges, C.S., Killgore, E. & Anderson, weeds of Australia. Inkata Press, Australia. 653 p. R.C. (1997). An evaluation of the rust fungus Rivals, P. (1960). Les espèces fruitières introduites Gymnoconia nitens as a potential biological agent à La Réunion (notes historiques et biologiques), for alien Rubus species in Hawaii. Biological Travaux du laboratoire forestier de Toulouse. Control 10, 151–158. Vol.1 art. 3. Université de Toulouse, Toulouse, Groves, R.H., Williams, J. & Corey, S. (1997) Towards France. an integrated management system for blackberry Robert, R. (1986). Climat et hydrologie de la (Rubus fruticosus L. agg.), Albury, New South Réunion, étude typologique, étude régionale Wales, Australia, pp. 151–199 de l’alimentation et de l’écoulement, Thèse Jacob De Cordemoy, E. (1895). Flore de l’île de la de doctorat, Université de Montpellier III, Réunion. Klincksieck, Paris, France. Montpellier, France. 438 p.

XIII International Symposium on Biological Control of Weeds - 2011 160 Session 4 Target and Agent Selection

Rohlf, F.J. (1993). NTYSYS-PC: Numerical plantes introduites à la Réunion et dynamique Taxonomy and Multivariate Analysis System, de la végétation sur les coulées volcaniques. Version 1.80. Exeter Publishers, Setauket, USA. Ecologie 26, 169–180. Sigala, P. & Lavergne, C. (1996) Environmental Thebaud, C. (1989). Contribution à l’étude des weeds in the native forests of La Reunion: plantes étrangères envahissantes à la Réunion. Prospects for biocontrol, In Proceedings of the. IX International Symposium on Biological IRAT-ONF, Saint Denis, Réunion. 49 p. Control of Weeds, (eds Moran, V.C. & Hoffman, Vaughan, R.E. (1937). Catalogue of the Flowering J.H.), pp. 339. Cape Town, South Africa Plants in the Herbarium Mauritius. Bulletin of Strasberg, D. (1995). Processus d’invasion par les the Mauritius Institute 1, 1–120.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 161

Beyond the Lottery Model: Challenges in the Selection of Target and Control Organisms for Biological Weed Control

P. B. McEvoy and K. M. Higgs

Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97331, USA [email protected]

Abstract

Biological control scientists have long tried to pick winning combinations of target and control organisms, while minimizing adverse effects on nontarget organisms. What have we learned from recent work and what should be focus on moving forward? Establishment rates for new control organisms are nearing 100% with little room for improvement; however, improving rates of control given establishment and minimizing off-target effects are two areas for improvement. Two ways to keep control organisms effectively on target are to (1) diagnose and exploit weed vulnerabilities using targeted life-cycle disruption (thereby achieving effective biological using fewer control organism individuals and species), and (2) investigate the mix of evolutionary and ecological forces enabling control organisms to exploit new habitats and hosts (thereby better forecasting outcomes of biological control). Continuity of program support, learning from experience, rational regulations and policies, plus developing and exchanging new technology will help achieve these goals.

XIII International Symposium on Biological Control of Weeds - 2011 162 Session 4 Target and Agent Selection

Bottom-Up Effects on Top-Down Regulation of a Floating Aquatic Plant by Two Weevil Species: The Context-Specific Nature of Biological Control

T. D. Center

Invasive Plant Research Laboratory, USDA-ARS, Fort Lauderdale, FL USA [email protected]

Abstract

Predicting the efficacy of prospective biological control agents, the holy grail of weed biological control, while often advocated, is rarely implemented. We examined, a posteriori, whether it would have been possible to predict which of two introduced weevil species, Neochetina eichhorniae Warner or N. bruchi Hustciche, would have been the superior choice for controlling Eichhornia crassipes (Mart.) Solms-Laubach. Plant nutrition and competition can alter a plant’s ability to sustain or compensate for herbivory and affect a phytophagous insect’s ability to reproduce. These factors could also influence efficacy predictions. We therefore conducted three outdoor mesocosm experiments to compare the performance of these two weevils, independently and together, among five fertilizer treatments. A low initial plant density experiment examined their ability to reduce growth and flowering but allowed for density to increase. A high plant density experiment evaluated their ability to lessen biomass and reduce surface coverage. The third experiment began with low plant density but plants were maintained at low density by harvesting a portion whenever coverage exceeded 50% of the water surface. This was intended to minimize intraspecific competition. The effects varied between weevil species, among fertilizer treatments, and among experiments. Interactions between herbivory and fertilizer treatments were apparent and the nature of these interactions varied among experiments. Efficacy therefore seemed nuanced and context specific, requiring extensive assessments of multiple evaluation criteria across a wide range of environmental and ecological conditions. Overly simplistic evaluations risk rejection of effective agents capable of mediating adverse impacts from invasive plant populations. These results also argue against the concept that a single best agent can be identified to control a weed that inhabits a broad range of habitats and conditions.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 163

Predicting Parasitism of Weed Biological Control Agents

Q. Paynter, S. V. Fowler, H. Gourlay, R. Groenteman, P. G. Peterson, L. Smith and C. J. Winks

Landcare Research, New Zealand [email protected]

Abstract

We conducted a nationwide survey of parasitism of weed biological control agents in New Zealand (NZ) and found that 19, mostly native, parasitoid species attack 10 weed biological control agent species. Fifteen of these parasitoid species were confined to five agents that possessed “ecological analogues”, defined as a native NZ insect that belongs to the same superfamily as the agent and occupies a similar niche on the target weed. Parasitoid species richness in NZ was positively correlated to richness in the area of origin. However, only agents with ecological analogues contributed significantly to this pattern. Our results support Lawton’s (1985) hypothesis that, to find enemy-free space, selected agents should “feed in a way that is different” and “be taxonomically distinct” from native herbivores in the introduced range.

XIII International Symposium on Biological Control of Weeds - 2011 164 Session 4 Target and Agent Selection

Learning from Experience: Two Weed Biological Control Programs with Rust Fungi Compared

L. Morin

CSIRO Ecosystem Sciences, GPO box 1700, Canberra, ACT 2601, Australia E-mail: [email protected]

Abstract

Rust fungi are the type of pathogens most widely used in classical biological control of weeds. This is primarily because of their typically high level of specificity, the severe damage they can inflict on plants, and efficient wind dispersal capability. They are not, however necessarily the easiest and most effective solutions to pursue for all weed problems. To illustrate the potential pitfalls that can be encountered with rust fungi as biological control agents, I will compare various aspects of the recent programs against bridal creeper (Asparagus asparagoides (L.) Druce) (Kleinjan et al., 2004; Morin and Edwards, 2006; Morin et al. 2002, 2006; Turner et al., 2010) and European blackberry (Rubus fruticosus L. aggregate) (Evans and Bruzzese, 2003; Evans et al., 2011; Gomez et al., 2008; Morin et al., 2011) in Australia. For example, specificity in rust fungi can be too high from a biological control point of view, necessitating the release of a range of rust pathotypes to affect the different genotypes of the weed that exist in the introduced range. Leaf-age resistance of the weed to the rust fungus can drastically limit its impact on individual plants, leading to only minor changes in the weed population dynamics. The growing season and preferred habitat of the target weed, as well as prevailing climatic conditions, can also influence the development of severe epidemics of the rust fungus and consequently its efficiency as a biological control agent.

References Kleinjan, C.A., Morin, L., Edwards, P.B. & Wood, A.R. (2004) Distribution, host range and Evans, K.J. & Bruzzese, E. (2003) Life history phenology of the rust fungus Puccinia myrsiphylli of Phragmidium violaceum in relation to its in South Africa. Australasian Plant Pathology 33, effectiveness as a biological control agent 263–271. of European blackberry. Australasian Plant Morin, L. & Edwards, P.B. (2006) Selection of Pathology 32, 231–239. biological control agents for bridal creeper: Evans, K.J., Gomez, D.R., Jones, M.K., Oakey, H. & a retrospective review. Australian Journal of Roush, R.T. (2011) Pathogenicity of Phragmidium Entomology 45, 287–291. violaceum isolates on European blackberry clones Morin, L., Willis, A.J., Armstrong, J. & Kriticos, and on leaves of different ages. Plant Pathology D. (2002) Spread, epidemic development and 60, 532–544. impacts of the bridal creeper rust in Australia: Gomez, D.R., Evans, K.J., Baker, J., Harvey, P.R. summary of results. In Proceedings of the 13th & Scott, E.S. (2008) Dynamics of introduced Australian Weeds Conference (eds Spafford Jacob, populations of Phragmidium violaceum H., Dodd, J. & Moore, J.H.) pp. 385–388. Plant and implications for biological control of Protection Society of Western Australia, Perth. European blackberry in Australia. Applied and Morin, L., Neave, M., Batchelor, K. & Reid, A. (2006) Environmental Microbiology 74, 5504–5510. Biological control: A promising tool for managing

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 165

bridal creeper, Asparagus asparagoides (L.) Druce, Australia. European Journal of Plant Pathology in Australia. Plant Protection Quarterly 21, 69– 131, 289–303. 77. Turner, P.J., Morin, L., Williams, D. & Kriticos, Morin, L., Evans, K.J., Jourdan, M., Gomez, D.R. D.J. (2010) Interactions between a leafhopper & Scott, J.K. (2011) Use of a trap garden to find and rust fungus on the invasive plant Asparagus additional genetically distinct isolates of the asparagoides in Australia: a case of two agents rust fungus Phragmidium violaceum to enhance being better than one for biological control. biological control of European blackberry in Biological Control 54, 322–330.

XIII International Symposium on Biological Control of Weeds - 2011 166 Session 4 Target and Agent Selection

Potential Benefits of Sourcing Biological Control Agents from a Weed’s Exotic Range

P. S y r e t t 1, R. Emberson2 and S. Neser3

1Landcare Research, PO Box 40, Lincoln 7640, New Zealand [email protected] 2Department of Ecology, PO Box 84, Lincoln University, Lincoln 7647, New Zealand [email protected] 3ARC-PPRI, Private Bag x134, Queenswood 0121, South Africa [email protected]

Abstract

Specialist herbivores that establish in the exotic range of a weed provide a potential source of biocontrol agents. Such agents are more likely to adapt successfully to another novel environment and to be devoid of parasitoids and disease. In addition, in a simplified system lacking many key phytophages, assessment of the impact of potential agents may be easier than in the native range. In a South African program for control of weedy Acacia species Bruchophagus acaciae (Cameron) has been collected from New Zealand where, having established without parasitoids, it is much more abundant than in its native Australia. New Zealand has accumulated many specialist herbivores of Acacia from Australia and one of these, the weevil, Storeus albosignatus Blackburn, might have provided a useful biocontrol agent for South Africa’s weedy Acacia species. This weevil was first recorded in New Zealand in the 1930s where it is now widespread and relatively common, feeding on seeds of several Acacia species. However, it was not associated with Acacia in Australia until 2009. Although seed-feeding agents of target Acacia species were sought in Australia intermittently over a period of more than 30 years and five species of Melanterius weevils were introduced to South Africa for control of different weedy Acacia species, the program did not collect S. albosignatus from Acacia species. Storeus albosignatus could possibly have provided a more parsimonious solution by limiting seed production of several of the target Acacia species. Searching for biocontrol agents in the exotic range will be most useful in areas where potential agents naturally accumulate in the exotic range through relative proximity to the native range, but it is also applicable in areas where establishment occurs through accidental transfer by human agency.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 167

Plant-Mediated Interactions among Herbivores: Considerations for Implementing Weed Biological Control Programs

L. R. Milbrath1 and J. R. Nechols2

1USDA-ARS, Ithaca, NY, USA 2Kansas State University, Manhattan, KS, USA [email protected]

Abstract

Complex trophic interactions are common, both in natural and managed ecosystems. One such interaction that has important implications for biological control of weeds involves plant responses to feeding by an herbivore which then impacts one or more other herbivores. Effects may be positive or negative, and mechanisms can be chemical or structural. Knowing if, and to what extent, these indirect plant-mediated interactions occur prior to importing new biological agents can assist with decisions about candidate selection, thus reducing economic and environmental costs, and increasing the overall success rate of weed biological control programs. We examined whether feeding by the musk thistle weevil Trichosirocalus horridus (Panzer), which attacks the vegetative crown early in the plant’s development, alters musk thistle as a resource for the later-arriving weevil Rhinocyllus conicus Frölich, which infests flower heads. Minor infestations of musk thistle by T. horridus had no effect on R. conicus oviposition and subsequent production of new adults. In contrast, heavy infestations of T. horridus reduced 1) R. conicus-musk thistle synchrony, 2) acceptability of musk thistle to ovipositing R. conicus, 3) the quantity and 4) the quality of resource available to R. conicus larvae. As a result, the production of new R. conicus adults was reduced 63%. Thus, even spatially- and temporally-isolated herbivores can affect one another negatively and in multiple ways. Nevertheless, musk thistle seed reduction was still greater when both weevils were present. Hence, the outcome for biological control programs may not necessarily be adverse because of compensatory trade-offs concerning the relative impacts of the two herbivores on the weed. Recommendations for incorporating protocols to assess potential indirect plant-mediated impacts on weed biological control programs will be given.

XIII International Symposium on Biological Control of Weeds - 2011 168 Session 4 Target and Agent Selection

The Use of Chemical Ecology to Improve Pre-Release and Post-Release Host Range Assessments for Potential and Released Biological Control Agents of Cynoglossum officinale

I. Park, M. Schwarzländer and S. E. Eigenbrode

University of Idaho, Moscow, ID USA [email protected]

Abstract

Unlike their extensive use in pest insect management systems, chemical ecological techniques are not commonly used in biological weed control programs, despite their potential benefits to improve host range predictions. Semiochemicals are the principal communication in insect-plant interactions and are especially important mediators of host selection behavior in response to different chemical cues. While much of pre-release host range testing focuses on differing test conditions to either assess the fundamental host range (no-choice) or the realized host range (choice and field tests), little research is directed at underlying host plant cues triggering or preventing a potential agent female’s host choice. This is true for two insects currently used or proposed as biological control agents for Cynoglossum officinale L., the root-mining weevil Mogulones cruciger Hbst. and the seed-feeding weevil, Mogulones borraginis (Fabricius). The former has been released for the control of C. officinale in Canada in 1997 but a pest alert has been issued for the insect by regulatory authorities in the USA in 2010 because of risks of non target plant feeding. M. borraginis, has a narrower host range and is being proposed for introduction into the USA. Pre-release host range evaluations for C. officinale agents are complicated by the fact that several native confamilials are rare and endangered or cannot be grown under greenhouse conditions. To assess the risk of non-target feeding by M. cruciger in the USA and predict the environmental safety of M. borraginis, we developed a portable system to collect headspace volatiles of native confamilials of C. officinale. We use combined gas chromatography and electroantennogram detection (GC-EAD) to test whether the volatile headspace or particular components of it, identified using gas chromatography with mass spectrometry (GC-MS), triggers responses of the antennas of either weevil species. We then verify preliminary GC-EAD results in behavioral trials using a four-area olfactometer. We assert that a major advantage of this approach is the non-destructive assessment of the attractiveness of rare non-target plant species to either weevil species.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 169

Shooting Straight: What Weeds Should We Target Next?

R. D. van Klinken

CSIRO Ecosystem Sciences, EcoSciences Precinct, PO Box 2583, Brisbane, Qld, Australia 4001 Email: [email protected]

Summary

Biocontrol is the only realistic option for managing some of our most serious weeds, but projects are typically long-term and costly to develop. Therefore it is critical that targets are objectively selected by assessing the likely return on investment. This can be difficult given that both stakeholders and biocontrol practitioners can have very different views on which weeds should be targeted. Further, the relative importance of a particular weed problem, and how feasible and realistic biocontrol is for the particular species, is frequently not considered. Despite these challenges, little attention has been given to the process of target selection within the biocontrol community.

Target prioritisation can be considered within a matrix that includes weed impact, the feasibility of biocontrol (the process constraints for the establishment and execution of a biocontrol program) and the likelihood of successful outcomes (for example the chance that biological control can mitigate the identified impact). Weed impact can be extremely difficult to evaluate and predict, and any assessment needs to consider the context in which the weed occurs, whether the weed is a passenger or driver of environmental change, what it is impacting (in terms of production, biodiversity and ecosystem services), and the temporal and spatial dimension of invasions. Assessment of feasibility includes an in-depth analysis of constraints and opportunities, ranging from the practical (e.g. access to the native- range), social or political (e.g. acceptance of biocontrol as a management solution), and ecological (e.g. the genetic integrity of the target within its native range and the availability of host-specific natural enemies). Likelihood of success includes determining whether biocontrol can address clearly defined criteria for the successful management of the weed, and whether weed impacts are likely to be mitigated by the available natural enemies.

Australian examples were used to illustrate how such a prioritisation matrix can help identify the next generation of biocontrol targets by (a) helping achieve the right balance between targeting weeds of moderate to high impact where potential agents are readily available and high impact weeds where they may not be, and (b) focusing research around key knowledge gaps preventing the adequate prioritisation of new and existing targets. The biocontrol “industry” needs to demonstrate continued success against important targets if it is to survive, and that this will require science-driven target selection, a balanced portfolio of targets, quantification of potential weed impact, and continued improvement in science and technology.

XIII International Symposium on Biological Control of Weeds - 2011 170 Session 4 Target and Agent Selection

Does Rise and Fall of Garlic Mustard Eliminate the Need for Biological Control?

B. Blossey1 and V. Nuzzo2

1Department of Natural Resources, Fernow Hall, Cornell University, Ithaca, NY 14853, USA [email protected] 2Natural Area Consultants, 1 West Hill School Road, Richford, NY 13835, USA

Abstract

Garlic mustard (Alliaria petiolata (M. Bieb.) Cavara and Grande), a European biennial herb introduced to North America 150 years ago is now widespread in temperate forests from East to West and as far north as Alaska and south to Georgia. Garlic mustard’s rapid spread, high local abundances, and lack of any beneficial effects spurred development of a biological control program in the late 1990’s, which included a long-term monitoring program to develop baseline data on life history and population dynamics of A. petiolata We anticipated that before and after comparisons would allow us to assess changes in A. petiolata performance associated with biological control agent releases. After a decade of studies in the East and Midwest, we have seen declines of A. petiolata to very low abundance at our permanent monitoring locations (but only in the absence of management efforts). Detailed evaluations of potential mechanisms point to negative soil feedback, i.e. the build- up of soil pathogens that appear to suppress A. petiolata while having no apparent negative effect on associated forest understory vegetation. While these effects appear widespread in the oldest invaded areas, similar population declines are not as widespread or apparent in the more recently invaded areas further west. We are now ready to petition the introduction of the first host specific control agent – but is this the right choice of action? At the present time the lack of clear and extensive negative ecosystem impacts of A. petiolata and the apparent decline without management intervention suggests that adding biological control agents to the North American fauna will not aid in restoration of herbaceous forest communities.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 171

Unravelling the Identity of Tamarix in South Africa and its Potential as a Target for Biological Control

M. Byrne, G. Mayonde and G. Goodman-Cron

School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, Wits 2050, South Africa [email protected]

Abstract

The Old World genus Tamarix has become naturalized and invasive in much of the rest of the world. Tamarix usneoides E. Mey. ex Bunge is native to south western Africa and indigenous to South Africa, where it is being used for phytoremediation of acid mine drainage from tailings storage facilities on gold mines. However, T. chinensis Lour., T. parviflora DC. and T. ramosissima Ledeb. are all exotic to South Africa, and are hypothesized to be hybridizing among themselves and with T. usneoides. Correct specific identification is therefore essential; to clone the indigenous species for phytoremediation use, and to investigate the possibility of biological control of the alien species. Tamarix remains one of the more taxonomically difficult genera to identify and when in the vegetative state many taxa are almost indistinguishable. The high incidence of hybridization in Tamarix also plays a role in the taxonomic confusion. The Internal Transcribed Spacer (ITS) regions of ribosomal DNA (rDNA) were successfully used to identify the local Tamarix species and their hybrids. Insect abundance and diversity were found to be higher on the indigenous T. usneoides than on the exotic T. ramosissima and its hybrids, suggesting that a potential biological control agent might distinguish between the alien and indigenous species. The potential for successful biological control of T. ramosissima in South Africa is discussed.

XIII International Symposium on Biological Control of Weeds - 2011 172 Session 4 Target and Agent Selection

Origins and Diversity of Rush Skeletonweed (Chondrilla juncea) from Three Continents

J. Gaskin1, C. L. Kinter2, M. Schwarzländer3, G. P. Markin4, S. Novak5 and J. F. Smith5

1USDA Agricultural Research Service, Northern Plains Agricultural Research Laboratory, 1500 N. Central Ave., Sidney, MT 59270 USA [email protected] 2Idaho Department of Fish and Game, Idaho Natural Heritage Program, 600 South Walnut, Boise, ID 83707 USA 3University of Idaho, Department of Plant, Soil and Entomological Sciences, Moscow, ID 83844 USA 4USDA Forest Service, Rocky Mountain Research Station, 1648 S. 7th Ave., Bozeman, MT 59717 USA 5Department of Biological Sciences, Boise State University Boise, ID 83725 USA

Abstract

Rush skeletonweed (Chondrilla juncea L.) is an invasive apomictic perennial plant in Australia, South- and North America, accidentally introduced from Eurasia, which shows differential resistance/tolerance to some herbicides and classical biological control agents. Rush skeletonweed biotypes have been locally described using morphology, phenology, isozyme patterns, and resistance to control agents, but studies comparing invasions on different continents and determining exact origins of invasive genotypes do not exist or are lacking in detail. Commonly available molecular tools and bulk analysis capacity now make it possible to determine genetic diversity within invasions and their origins. We investigated over 1000 plants from three invaded continents using highly variable AFLP (Amplified Fragment Length Polymorphism) markers, and found 13 distinct genotypes (three from Australia, three from Argentina, and seven from North America). No genotypes were shared between continents. Certain regions in North America, such as California, contain only one genotype of the weed. We then investigated over 1000 plants from the native Eurasian range to determine, as accurately as possible, origins of the invasive genotypes, including those that are currently resistant to strains of rust (Puccinia chondrillina Bubak & Syd.) used in biological control programs. This information can be used to screen for pathogens and other agents that will not be resisted or tolerated by certain rush skeletonweed genotypes. Understanding global intraspecific diversity and exact origins can improve management of differentially-resistant/tolerant weed biotypes, enhance efficacy of future agent selection, and increase cooperation between invaded regions.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 173

Comparing the Population Biology of Isatis tinctoria in its Native Eurasian and Introduced North American Range under Different Experimental Treatments

R. Gibson1, H. L. Hinz2 and M. Schwarzländer1

1University of ID, Mosco, USA [email protected] 2CABI Europe-Switzerland, Delemont, Switzerland

Abstract

Isatis tinctoria L. is an annual, biennial, or short-lived perennial mustard native to Eurasia. The weed was culturally introduced to North America where it has naturalized and become a noxious weed in several western US states. The goals of this study are to identify the sensitive phenostages of I. tinctoria to improve management strategies for the invasive plant and to identify most effective biological control agents. Permanent study plots were established in 2010 at an experimental and a close by natural infestation of I. tinctoria in southeastern Idaho. In Europe, the experimental site was established in 2008 in southern Germany. The study site in Germany includes the following treatments: Presence and absence of 1) interspecific competition and 2) biological control insect herbivory. The experimental site in Idaho includes presence and absence of 1) interspecific competition and 2) the native rust pathogen (Puccinia thlaspeos Schub.), as treatments. Established with each experimental site were seed bank plots with presence and absence of interspecific competition and seed longevity experiments. In both ranges, interspecific competition had so far the strongest and most consistent effect on plant numbers. In Europe, specific insect herbivores negatively impacted individual plant parameters. We hope that the continuation of the study will allow us to demonstrate effects of treatments on life stages of I. tinctoria. Results will ideally help prioritizing candidate insect biological control agents currently studied in the native range and assessing the net effect of the native rust present in the introduced range.

XIII International Symposium on Biological Control of Weeds - 2011 174 Session 4 Target and Agent Selection

Invasive Exotic Plant Species in Tennessee, USA: Potential Targets for Biological Control

J. Grant, G. Wiggins and P. Lambdin

Department of Entomology and Plant Pathology, 205 Ellington Plant Sciences Building, The University of Tennessee, Knoxville, TN 37996 USA [email protected]

Abstract

Numerous invasive exotic plant species are well established in Tennessee, located in the southeastern United States, where many of these plant species pose serious economical and environmental threats to agriculture, forestry, natural areas, and urban areas. The threat of some invasive plant species, such as musk thistle (Carduus nutans L.), multiflora rose (Rosa multiflora L.), and kudzu (Pueraria montana var. lobata (Willd.) Maesen and S. Almeida), is well documented, and these species are well recognized by growers and landowners. Management programs, primarily focused on chemical, mechanical and cultural controls, have been developed to limit the spread of these exotic plant species, and biological control is a major component of integrated pest management of musk thistle. However, the threat of other species, such as purple loosestrife (Lythrum salicaria L.), spotted knapweed (Centaurea stoebe L. subsp. micranthos), and Canada thistle (Cirsium arvense (L.) Scop.), in Tennessee is not as well known, and growers and landowners are not as familiar with these exotic species. In most cases, the state-wide distribution of many of these invasive plant species has not been clearly defined. In addition, the diversity of native and exotic insect herbivores that utilize many of these introduced plant species has not been investigated. Biological control will be explored as a management tactic against selected exotic weeds based upon the ‘best fit’ for the climate and geography of Tennessee. This poster details invasive exotic plant species established in Tennessee and examines the potential for the integration of biological control into pest management programs directed against selected weed species.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 175

Genetic Variation in a Biological Control Target Weed: The Strawberry Guava Species Complex

P. Johansen1, R. Manshardt1 and T. Johnson2

1Tropical Plant & Soil Sciences Dept., University of Hawaii at Manoa, Honolulu, HI USA 2Institute of Pacific Island Forestry, Pacific SW Res. Sta., USDA Forest Service, Volcano, HI USA

Abstract

Our objective is to characterize the genetic variation in strawberry guava populations in Hawaii, with the goal to inform biological control efforts currently being developed to counter this invasive species complex in native forests. Specimens collected on the islands of Hawaii, Maui, and Oahu were evaluated for fruit and vegetative morphology, ploidy as determined by flow cytometry, and microsatellite variation at three chloroplast SSR loci and three nuclear SSR loci. Results supported three previously recognized taxa and one new category. Psidium littorale Raddi was uniform with regard to fruit morphology (yellow, spindle-shaped), ploidy (8x), and SSR polymorphisms, suggesting that it may be a fertile allo-octoploid or a sterile apomict. Similarly, P. lucidum Hort. was uniform in fruit morphology (yellow, spherical fruits) and SSR genotype, but showed minor ploidy variation about 6x, suggesting a mostly fertile allo-hexaploid or an apomict with some residual sexual function. P. cattleianum Afzel. ex Sabine displayed a single uniform chloroplast and nuclear SSR genotype, but ploidy variation between 6.5x and 7.1x, and red fruit color of variable hue and intensity, suggesting that sexual reproduction is operative in this nominally heptaploid form and that it produces mainly aneuploid progeny. A fourth form (Psidium “X”) with fruit color (orange) and ploidy range (6.4x to 6.8x) intermediate between those of P. lucidum and P. cattleianum originally suggested derivation through interspecific sexual crossing or possibly elimination of genetic material in the aneuploid sexual progeny of hybrids or of self- or sib-mated P. cattleianum. However, the presence of a unique chloroplast SSR allele found in the orange-fruited forms and not in either of the putative parent species indicates that it is not recently of hybrid origin or directly derived from P. cattleianum. The orange-fruited form represents a new taxon not previously described in Hawaii. The SSR uniformity within the four strawberry guava taxa may reflect predominantly apomictic seed production, or simply that our survey employed an inadequate number of marker loci to detect polymorphisms. This apparently modest level of genetic variation may suit the strawberry guava complex in Hawaii to target status for a host-specific biological control agent, such as Tectococcus ovatus Hempel.

XIII International Symposium on Biological Control of Weeds - 2011 176 Session 4 Target and Agent Selection

Demographic Matrix Model for Swallow-Wort (Vincetoxicum spp.)

L. R. Milbrath1 and A. S. Davis2

1USDA-ARS, Ithaca, NY, USA [email protected] 2USDA-ARS, Urbana, IL, USA

Abstract

Demographic matrix modeling of plant populations can be a powerful tool to identify key life stage transitions that contribute the most to population growth of an invasive plant and hence should be targeted for disruption (weak links) by biological control and/or other control tactics. Therefore, this approach has the potential to guide the selection of effective biological control agents. We are in the process of parameterizing a five life-stage matrix model in order to generate pre-release agent recommendations for the swallow-wort biological control program. Pale swallow-wort (Vincetoxicum rossicum (Kleo.) Barb.) and black swallow-wort (V. nigrum L.) are herbaceous, perennial, viny milkweeds introduced from Europe (Apocynaceae-subfamily Asclepiadoideae). Both species are becoming increasingly invasive in a variety of natural and managed habitats in the northeastern United States and southeastern Canada. Black swallow-wort appears restricted to higher light environments, whereas pale swallow-wort infestations occur from the high light environments of open fields to low light forest understories. We are quantifying demographic transitions over 3-4 years of both swallow-wort species in field and, for pale swallow-wort, forest habitats in New York State (N = six populations). Vital rates estimated include seed survival, germination, plant survival to reproductive maturity, and fecundity (viable seeds produced per plant). Data will be presented on model parameters derived to date.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 177

How Many Species of Salsola tumbleweeds (Russian Thistle) Occur in the Western USA?

L. Smith1, G. F. Hrusa2 and J. F. Gaskin3

1USDA-ARS-WRRC, 800 Buchanan Street, Albany, CA 94710, USA [email protected] 2California Department of Food and Agriculture-PHPPS, Botany Laboratory, 3294 Meadowview Rd., Sacramento, CA 95832, USA 3USDA-ARS-NPARL, 1500 N. Central Avenue, Sidney, MT 59270, USA

Abstract

Russian thistle or common tumbleweed, Salsola tragus L. (sensu lato), is an alien weedy annual plant that infests over 41 million hectares in the western United States. The taxonomy of this plant has had a long confusing history, with frequent misapplication of the species names kali and australis. Recent studies based on morphology, allozymes and molecular genetics indicate that “Russian thistle” comprises seven distinct species in North America. Salsola tragus is probably the most widespread species. Salsola collina Pall. occurs primarily east of the Rocky Mountains, S. paulsenii Litv. primarily in deserts, and S. kali is restricted to ocean shores and is not a rangeland weed. Salsola australis R., sometimes reported as “type B”, occurs primarily in California, South Africa and Australia, but has never been documented to occur in Eurasia. Almost all uses of this name before 2008 are probably misapplications. Polyploid hybrids include S. x gobicola (includes S. tragus and S. paulsenii ancestry), which is known from western USA and central Asia, and S. x ryanii (includes S. tragus and S. australis ancestry), which is known only from California. A gall forming midge, Desertovelum stackelbergi Mamaev, from Uzbekistan (Sobhian et al. 2003. Biol. Control 28: 222-228) and a fungal pathogen, Colletotrichum gloeosporoides (Penz), from Hungary (Bruckart et al. 2004. Biol. Control 30: 306-311) had much higher rates of attack and damage to S. tragus than S. australis. Although it was previously believed that all species in the kali section of Salsola originated in Eurasia, the presence of 4 indigenous species in Australia suggests a separate clade (Borger et al. 2008. Aust. J. Bot. 56: 600–608). It is likely that S. australis is native to Australia.

XIII International Symposium on Biological Control of Weeds - 2011 178 Session 4 Target and Agent Selection

An Initial Focus on Biological Control Agents for the Forest Invasive Species Prosopis juliflora in the Dry Zone of Myanmar

W. W. Than

Senior Researcher, Forest Protection Service, Forest Research Institute, Yezin, Myanmar; National Focal Point of Asia-Pacific Forest Invasive Species Network [email protected]

Abstract

Prosopis juliflora (Sw.) DC (Prosopis hereafter) was introduced into Myanmar for dry zone greening about 1950 from Israel by Agriculture Research and Development Cooperation (ARDC). Prosopis is an aggressive weed of forests in the dry zone of Myanmar. Though regarded useful for fuel wood for rural people, this deep rooted species forms thickets, outcompetes native vegetation, has hard and sharp thorns. It is also capable of vegetative reproduction and produces many coppices when cut, and is hence difficult to eradicate. Its distribution is expanding and control is needed in some areas. Investigation for biological control agents were conducted in Pyawbwe in January, 2010. Prosopis appeared to struggle to compete with the climber Combretum roxburghii D. Don, and the shrubs Azima sarmentosa Benth. and Lantana camara L. The use of these plants as competitors to suppress Prosopis growth is not desirable as Prosopis has benefits for rural people, while these competitors do not. Damage to Prosopis was detected in the form of yellowing foliage and damage from pathogens around the cuts made to the woody tissue during harvest of fuel wood. Fusarium sp., Tubercularia sp. and Nectria sp. were identified from these damaged trees. Small-scale trials have been initiated to examine the potential for these fungal pathogens to aid in the biological control of Prosopis.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 179

Potential for the Biological Control of Crassula helmsii in the U.K.

S. Varia and R. Shaw

CABI, Bakeham Lane, Egham, Surrey, TW20 9TY, UK [email protected]

Abstract

Crassula helmsii (Kirk) Cockayne, also known as Australian swamp stonecrop or New Zealand pigmyweed, is native to Australia and New Zealand as the common names suggest. Since being introduced to the UK in 1911 as an ‘oxygenating’ plant for garden ponds, it has been gradually increasing its range both in the U.K. and parts of Europe by escaping gardens and through incorrect disposal by aquarium and pond owners. It has now spread to at least 2000 sites in the UK, particularly threatening conservation sites that are home to rare and endangered organisms. With no dormant period and a high tolerance to a range of temperatures, it can dominate static and slow moving water bodies, as well as bank sides, growing in dense mats as an emergent, submerged or terrestrial form. Once established its impacts can be serious; affecting native biodiversity and impeding water flow. With limited possibilities for chemical control in the EU and the plant’s ability to re-grow from fragments as small as 1cm, this weed is particularly difficult to manage. An estimation of the cost of treating 2000 infested sites was estimated to be between €5.8 - 12 million. Little is known of the natural enemy complex on C. helmsii populations in the native range, so CABI and collaborators initiated scoping surveys in New Zealand and Australia in 2009. Despite the limited nature of the initial surveys, considerable pathogen and herbivore damage were observed, revealing an assortment of natural enemies associated with this weed. The discovery of two highly damaging stem-mining weevils that were previously unrecorded, suggests more species may be identified in the native range in following future surveys. The discovery of additional natural enemies, to those specified in the literature, bodes well for the future biological control of this species.

XIII International Symposium on Biological Control of Weeds - 2011 180 Session 4 Target and Agent Selection

The Road Less Taken: A Classical Biological Control Project Operated Through an NGO

A. McClay1, M. Chandler2, H. L. Hinz3, A. Gassmann3, V. Battiste4 and J. Littlefield5

1McClay Ecoscience, 15 Greenbriar Crescent, Sherwood Park, Alberta, Canada T8H 1H8 Email: [email protected] 2 Minnesota Dept of Agriculture, 625 North Robert Street, Saint Paul, MN 55155, USA Email [email protected] 3CABI-Europe Switzerland, 1, rue des Grillons, Delémont CH-2800, Switzerland Email: [email protected], [email protected] 4Alberta Invasive Plants Council, PO Box 869, Okotoks, Alberta, Canada T1S 1A9 Email: [email protected] 5Department of Entomology, Montana State University, PO Box 173020, Bozeman, MT 59717- 3020 USA Email: [email protected]

Abstract

Historically, most classical weed biological control projects have been initiated and managed by government departments or agencies. We have successfully developed a project along an alternative route, where the main sponsor is a non-governmental organization. The Alberta Invasive Plants Council (AIPC) was established in 2006 to raise the awareness of invasive plants as a problem in Alberta and to promote cooperative efforts to manage these plants in the province. Later that year, AIPC obtained funding to start a biological control project against common tansy (Tanacetum vulgare L.), a toxic, perennial weed native to Europe and Asia which is spreading rapidly in pastures and riparian areas in western Canada and the northern USA. This is a joint US-Canadian project, with funding from a number of state, provincial, federal, and industry sponsors in both countries. The US funding is coordinated by the Minnesota Department of Agriculture, and most Canadian funding is handled through AIPC. The CABI laboratory in Delémont, Switzerland, conducts exploration and agent testing under contract with AIPC, and a private consultant (the first author) coordinates the project, also under contract with AIPC. The project coordinator prepares proposals for ongoing funding, reports on progress to project sponsors, promotes public awareness of the project, and works with CABI to select agents for study and oversee the project. AIPC provides administrative services to the project including contract management, billing, and accounting. An annual consortium meeting is held to review progress and plan future work. This approach has allowed us to enhance the existing Canadian capacity for managing overseas biological control research.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 181

A Reassessment of the Use of Plant Pathogens for Classical Biological Control of Tradescantia fluminensis in New Zealand

D. M. Macedo1, O. P. Liparini1, R. W. Barreto1, and N. Waipara2

1Departamento de Fitopatologia, Universidade Federal de Viçosa, 36570 - 000, Viçosa, Minas Gerais, Brazil [email protected] 2Landcare Research, Private Bag 92170, Auckland, New Zealand

Abstract

Tradescantia fluminensis Vell. is an herbaceous plant of the family Commelinaceae, native to South America, and commonly known as small-leaf spiderwort. In 1910, it was introduced into New Zealand and became an aggressive invasive weed in natural ecosystems, causing serious environmental imbalances: in particular, preventing the regeneration of native species, and thus undermining the local biodiversity. The problem caused by T. fluminensis quickly led to the use of conventional control measures, but without success. As a result, biological control is considered to be the most promising alternative to mitigate the problems generated by T. fluminensis. Since 2003, surveys have been conducted in search of plant pathogens in south and southeastern Brazil. In the first phase of surveys, five new fungal species were added to the known pathogenic mycobiota of T. fluminensis: Ceratobasidium sp., Cercospora apii Fresen., Kordyana sp., Uredo sp. and Mycosphaerella sp. In 2009, a second phase of surveys was performed, and resulted in the addition of four more species of pathogenic fungi to the list: Colletotrichum falcatum Went, Rhizoctonia solani Kuhn, Sclerotium rolfsii Sacc. and Septoria sp. During this period, Kordyana (Brachybasidiaceae = Exobasidiales) - commonly associated with species of Commelinaceae, causing symptoms equivalent to those of white smuts such as Entyloma ageratinae Barreto and eVans, the successful biological control agent of Ageratina riparia in Hawaii and New Zealand, was assessed to have the most potential, based on field data. Subsequently, pathogenicity tests were conducted to evaluate its potential use as a biological control agent. Besides confirmation of its infectivity to T. fluminensis, its specificity to this target species was also demonstrated, among the 70 plant species that were tested. Based on these results, further studies are in progress to clarify the identity of Kordyana sp. and to develop a protocol for handling the introduction of the pathogen into New Zealand.

XIII International Symposium on Biological Control of Weeds - 2011 182 Session 4 Target and Agent Selection

European Insects as Potential Biological Control Agents for Common Tansy (Tanacetum vulgare) in Canada and the United States

A. Gassmann1, A. McClay2, M. Chandler3, J. Gaskin4, V. Wolf1,5 and B. Clasen6

1CABI-Europe Switzerland, 1, rue des Grillons, Delémont CH-2800, Switzerland Email: [email protected] 2McClay Ecoscience, 15 Greenbriar Crescent, Sherwood Park, Alberta, Canada T8H 1H8 Email: [email protected] 3Minnesota Dept of Agriculture, 625 North Robert Street, Saint Paul, MN 55155, USA Email: [email protected] 4USDA-ARS Northern Plains Agricultural Research Laboratory, 1500 N. Central Ave. Sidney, MT USA 59270 Email: [email protected] 5University of Bielefeld, Universitätsstraße 25, D-33615 Bielefeld, Germany Email: [email protected] 6Department of Horticultural Science, University of Minnesota, 326 Alderman Hall, 1970 Folwell Ave., St. Paul, MN 55108 , USA Email: [email protected]

Abstract

Common tansy (Tanacetum vulgare L., Asteraceae), an herbaceous perennial native to Europe, was introduced into North America as a culinary and medicinal herb. Now widely naturalized in pastures, roadsides, waste places, and riparian areas across Canada and the northern USA, tansy is also spreading in forested areas. It contains several compounds toxic to humans and livestock if consumed, particularly α-thujone, and is listed as a noxious weed in several states and provinces. A biological control program for common tansy is being coordinated by a Canadian-US consortium led by the Alberta Invasive Plant Council and the Minnesota Department of Agriculture, with CABI Switzerland Centre identifying and testing potential agents for efficacy and host specificity. Collection efforts are focused on Eastern Europe (Russia and Ukraine) to maximize the climatic match with the infested areas in North America. Several potential agents are under study, the most promising agent at present being a stem-mining weevil, Microplontus millefolii (Schltz.). A root-feeding flea beetle, Longitarsus noricus Leonardi, also shows promise, and DNA barcoding is being used to separate this species from morphologically similar species that may emerge as contaminants in host-specificity tests. The leaf-feeding tortoise beetle Cassida stigmatica Suffr. is specific to Tanacetum but is able to complete development on the North American native T. bipinnatum ssp. huronense (Nutt.) Breitung; further evaluation of the risk to this species is needed. Life history studies on a stem-mining moth, Isophrictis striatella (Denis & Schiffermüller), suggest that it develops mainly in the previous year’s dead stems. This may reduce its potential impact as a biological control agent. The effects of chemical and genetic variation in tansy on the feeding and oviposition responses of insects are being studied, and molecular methods are also being used to evaluate the relationships between T. vulgare and other species.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 183

The Potential for the Biological Control of Himalayan Balsam Using the Rust Pathogen Puccinia cf. komarovii: Opportunities for Europe and North America

R. Tanner1, C. Ellison1, H. Evans1, Z. Bereczky2, E. Kassai-Jager2, L. Kiss2, G. Kovacs 2, 3 and S. Varia1

1CABI, Bakeham Lane, Egham, Surrey, TW20 9TY, UK [email protected] 2Plant Protection Institute of the Hungarian Academy of Sciences, H-1525 Budapest, P.O. Box 102, Hungary 3Eotvos Lorand University, Dept. Plant Anatomy, Budapest, Hungary

Abstract

Himalayan balsam (Impatiens glandulifera Royle) is a highly invasive annual herb, native to the western Himalayas, which has spread rapidly throughout Europe, Canada and the United States since its introduction as a garden ornamental. The plant can rapidly colonize riparian systems, damp woodlands and waste ground where it reduces native plant diversity, retards woodland regeneration, outcompetes native plants for space, light and pollinators and increase the risk of flooding. Current control methods are fraught with problems and often unsuccessful due to the need to control the plant on a catchment scale. Since 2006, CABI and our collaborators have surveyed populations of Himalayan balsam throughout the plants native range (the foothills of the Himalayas, Pakistan and India) where numerous natural enemies have been collected and identified. Agent prioritization, through field observations and host range testing has narrowed the potential candidates down to the rust pathogen, Puccinia cf komarovii Tranzschel. This autoecious, monocyclic pathogen shows great promise, not only due to its impact on the host but also due to its high specificity as observed in the field and preliminary host range testing. The aecial stage infects the hypocotyl of young seedlings as they germinate through leaf litter containing teliospores. This initial infection severely warps the structure of the developing plant. The aeciospores then infect developing leaves to produce the cycling phase (uredia). This severely affects the photosynthetic capacity of the maturing plant, with the potential of reducing seed-set. Late in the season, teliospores are produced which overwinter in the leaf litter. This paper will review the research conducted to-date, including a molecular comparison of P. cf komarovii with other closely related species, the life cycle and infection parameters of the rust and an up-date on the current host specificity testing under quarantine conditions in the UK.

XIII International Symposium on Biological Control of Weeds - 2011 184 Session 4 Target and Agent Selection

The Scotch Broom Gall Mite: Accidental Introduction to Classical Biological Control Agent?

J. Andreas1, T. Wax1, E. Coombs2, J. Gaskin3, G. Markin4 and S. Sing4

1Washington State University, Puyallup WA, USA [email protected] 2Oregon Department of Agriculture, Salem OR, USA 3USDA ARS NPARL, Sidney MT, USA 4USFS Rocky Mountain Research Station, Bozeman MT, USA

Abstract

The gall mite, Aceria genistae (Nal.) Castagnoli s.l., an accidentally introduced natural enemy of Scotch broom (Cytisus scoparius (L.) Link), was first discovered in the Portland OR and Tacoma WA region in 2005. It has since been reported from southern British Columbia to southern Oregon. Observationally, the mite appears to reduce Scotch broom seed production and at high densities can cause extensive stem die-back and plant mortality. In order to utilize the mite as a classical biological control agent, a study of its host range and potential nontarget attack was started in 2006 and continued in 2008- 2010. Over 20 ecologically- and economically-valued species were tested in greenhouse and open-field studies. Surveys of confamilial nontarget plant species naturally co- occurring with mite-infested Scotch broom were also assessed. Mites and gall formations were noted on hybrids and ornamental species of Scotch broom. Under greenhouse tests, gall-like growth and eriophyid mites were found on Lupinus densiflorus Benth. (L. microcarpus), a species listed as endangered in Canada. One unidentified eriophyid mite and no deformed growth was detected on L. densiflorus at naturally occurring populations growing sympatrically with mite-infested Scotch broom on Vancouver Island. The ambiguous taxonomy of the mites found on Scotch broom, gorse (Ulex europaeus L.) and L. densiflorus has added further complications to the study. The overall project and plans for developing a petition for its approval as a biological control agent will be discussed.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 185

The Impact of the Milfoil Weevil Eubrychius velutus on the Growth of Myriophyllum spicatum and Other Watermilfoils Native to Europe

J.-R. Baars

BioControl Research Unit, School of Biology and Environmental Science, University College Dublin, Ireland [email protected]

Abstract

Eurasian milfoil Myriophyllum spicatum L. (Haloragaceae) is an aquatic submerged macrophyte that has become invasive in several countries outside its native range. In the United States a native weevil Euhrychiopsis lecontei (Dietz) (Coleoptera: Curculionidae) is being used as a natural enemy of M. spicatum. In Europe, a similar aquatic weevil Eubrychius velutus Beck develops on several native milfoil species. Impact studies were conducted on three milfoil species including M. spicatum to assess the growth impact of the largely leaf-feeding habit of E. velutus. The results indicate that the adult and larval damage significantly reduces the growth rate of plants, even at low beetle densities. Although it may not be specific enough to release in regions that have closely related native milfoils, E. velutus may be utilized as a candidate agent elsewhere in the world. The implications of the levels of damage are discussed and compared with the damage characteristics of E. lecontei, informing the potential use of the weevil as a classical biological agent.

XIII International Symposium on Biological Control of Weeds - 2011 186 Session 4 Target and Agent Selection

Field Explorations in Anatolia for the Selection of Specific Biological Control Agents for Onopordum acanthium (Asteraceae)

M. Cristofaro1, F. Lecce2, A. Paolini2, F. Di Cristina2, L. Gültekin3 and L. Smith4

1ENEA C.R.Casaccia, UTAGRI-ECO, Via Anguillarese, 301, 00123 S. Maria di Galeria (Rome), Italy [email protected] 2Biotechnology and Biological Control Agency, Via del Bosco, 10, 00060 Sacrofano (Rome), Italy 3Atatürk University, Faculty of Agriculture, Plant Protection Department, 25240 TR Erzurum, Turkey 4 USDA-ARS, 800 Buchanan Street, Albany, CA 94710, USA

Abstract

Scotch thistle, Onopordum acanthium L. (Asteraceae), is a biannual thistle of Eurasian origin, accidentally introduced in North America in the late 1800s. It occurs in most of the western states especially in rangelands, pastures and disturbed soils. When abundant, it reduces forage availability for livestock and wildlife. Western Europe has previously been explored for prospective biological control agents, and seven species of insects have been released in Australia. However, host specificity requirements in N. America are more stringent for these agents because of the presence of many native Cirsium species. Some of the agents released in Australia do not appear to be sufficiently specific to use in N. America. Therefore we decided to conduct foreign exploration further east in areas not previously explored. Since 2007 we conducted several explorations and survey trips to discover new potential biological control agents in Turkey. Among the most promising candidates are three weevils Larinus latus (Herbst.), L. gigas Petri and L. grisescens Gyllenhal that were collected in Central Anatolia. The larvae of these species develop in the flowerheads and destroy most seeds. Other candidates are Psylliodes cf. chalcomera Illiger and Lixus cardui Olivier, whose larvae develop inside stems and leaves. Preliminary laboratory host specificity tests show a narrow host range of some biotypes of both potential biological control agents. Specimens from these experiments are currently undergoing genetic and morphological study to understand if there are distinct genetic entities not distinguishable by morphological traits. A new species of eriophyoid mite, Aceria sp., was recorded for the first time in the vicinity of Isparta, Western Turkey associated with the target weed. It causes stem atrophy and flower bud abortion with a consequent decrease of seed production. An unidentified nematode species causing visible blisters on Scotch thistle leaves was found in Central Turkey.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 187

Potential Biological Control of Invasive Tree-of-Heaven (Ailanthus altissima)

D. D. Davis and M. T. Kasson

Department of Plant Pathology, Penn State University, University Park, PA 16802, USA [email protected]

Abstract

The highly invasive tree-of-heaven, Ailanthus altissima (Mill.) Swingle, was introduced into Pennsylvania (PA), USA, in 1784 and has since spread across PA and most of the USA. Wherever it is found, Ailanthus often dominates a site at the expense of native plant species. However, in 2002-2003, we discovered several stands of dead and dying Ailanthus trees within oak-dominated, mixed-hardwood forests in south- central PA. Isolations from symptomatic Ailanthus seedlings, root sprouts, saplings, and canopy trees in the field consistently yielded the naturally occurring, soil- borne, wilt fungus Verticillium albo-atrum Reinke & Berthold. Identification was based on morphological characteristics, and confirmed using molecular techniques. Potted Ailanthus seedlings in the greenhouse and canopy Ailanthus trees in the field were inoculated with a randomly selected isolate of V. albo-atrum. Classic wilt symptoms quickly developed on inoculated plants, from which V. albo-atrum was recovered, fulfilling Koch’s Postulates and illustrating that V. albo-atrum was highly virulent on Ailanthus. As of 2011, V. albo-atrum has killed thousands of canopy Ailanthus trees in south-central PA. Intermingled non-Ailanthus tree species and understory shrubs have been generally unharmed. We hypothesize that a naturally occurring strain of V. albo-atrum has become host-adapted to Ailanthus. As part of risk assessment, we inoculated more than 80 non-Ailanthus species (potted seedlings in the greenhouse, as well as trees in the field) with a randomly selected strain of V. albo-atrum. Inoculated non-Ailanthus species were generally unharmed, again indicating that a pathogenic strain of V. albo-atrum may have become host-adapted to Ailanthus. In addition, past forest management practices in the area (e.g., clear-cutting large blocks of oaks killed by insect infestations) favored development of dense, nearly monoculture stands of clonal Ailanthus, which in turn. Short-range dissemination of V. albo-atrum within infected stands of Ailanthus likely occurs via intraspecific root grafts between diseased and healthy trees within these dense stands. Long-range dissemination of V. albo-atrum may be facilitated by Euwallacea validus (Eichhoff), an introduced ambrosia beetle that is epidemic on Ailanthus trees under stress from V. albo-atrum infections.

XIII International Symposium on Biological Control of Weeds - 2011 188 Session 4 Target and Agent Selection

Abrostola clarissa (Lepidoptera: Noctuidae), a New Potential Biological Control Agent for Invasive Swallow-Worts, Vincetoxicum rossicum and V. nig r um

M. Dolgovskaya1, M. Volkovitsh1, S. Reznik1, V. Zaitzev1, R. Sforza2 and L. Milbrath3

1Zoological Institute RAS, 199034, St.Petersburg, Russia, Russian Biocontrol Group [email protected] 2USDA-ARS-EBCL, Montferrier sur Lez, France 3USDA-ARS, Ithaca, New York, USA

Abstract

Pale and black swallow-worts (Vincetoxicum rossicum (Kleopow) Barbar. and V. nigrum (L.) Moench; Apocynaceae, subfamily Asclepiadoideae), perennial vines native to Eurasia, are now invading natural and anthropogenic habitats in the northeastern U.S.A. and southeastern Canada, threatening natural biodiversity and increasing control costs for land managers. Chemical and mechanical methods have not been adequate to control swallow- worts. In addition, no local American herbivores or pathogens cause significant damage to these weeds. Several potential biological control agents associated with Vincetoxicum spp. in Europe have been found and investigated, but none of them have yet been introduced. During explorations for herbivorous insects feeding on Vincetoxicum species in the Russian North Caucasus, we discovered a new potential biological control agent, Abrostola clarissa Staudinger (Lepidoptera, Noctuidae). A. clarissa inhabits low mountains and dry hills, having 1 – 2 generations per season. The biology of this species is similar to that of the closely related A. asclepiadis: eggs are laid on the undersurface of the host plant leaves, and larvae feed on the foliage and pupate in the soil. In natural conditions, larvae of this noctuid moth were collected only on Vincetoxicum spp. No-choice tests conducted under laboratory conditions showed that larvae of A. clarissa voluntarily fed and successfully pupated on Vincetoxicum nigrum, V. rossicum, V. hirundinaria (L.) Pers., and V. laxum (Bartl.) C.Koch. Neither feeding nor survival was recorded on other Apocynaceae (11 species of Amsonia, Apocynum, Asclepias, and Cynanchum) or on plants from other, more distantly related, families (Rubiaceae, Scrophulariaceae, and Convolvulaceae). We conclude that A. clarissa can be considered a highly specific potential biological control agent that undoubtedly deserves further study.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 189

Suitability of Using Introduced Hydrellia spp. for Management of Monoecious Hydrilla verticillata

M. J. Grodowitz1, J. G. Nachtrieb2, N. E. Harms1 and J. E. Freedman1

1US Army Engineer Research and Development Center, Vicksburg, MS, USA [email protected] 2US Army Engineer Research and Development Center, Lewisville, TX USA

Abstract

Two species of introduced leaf-mining flies, Hydrellia pakistanae Deonier and H. balciunasi Bock, suppress dioecious hydrilla by reducing photosynthesis, thereby impacting biomass production, tuber production, and fragment viability. To determine the flies’ suitability for monoecious hydrilla management several studies were conducted. When reared on monoecious hydrilla, H. pakistanae survival was reduced by 40 and 26 percent during bioassays and greenhouse rearing, respectively. Hydrellia spp. also exhibited a 9 day increase in developmental time when reared on monoecious hydrilla. Hydrellia spp. colonization and percent leaf damage between biotypes also differed significantly using small-scale tank experimentation. Initial fly stocking rates were equal, but four weeks after release fly levels and percent leaf damage in dioecious tanks were 5.3 and 2.4 fold higher than on monoecious. Small pond experimentation revealed similar results with fly levels 50-fold higher on dioecious in comparison to monoecious. Most importantly, release of close to 2,000,000 Hydrellia spp. at sites on Lake Gaston, NC since 2004 have failed to provide convincing evidence of establishment let alone population increase and impact. Recent anecdotal information from Guntersville Reservoir, AL suggests that shifts from a dioecious hydrilla dominated system to one composed mainly of the monoecious biotype may have been, in part, caused by differential feeding by Hydrellia spp. Experiments and field studies conducted since 2004 indicate that the monoecious biotype is not as suitable a host for introduced Hydrellia spp. as is dioecious hydrilla. This conclusion is based in part on reduced survival and longer developmental time in bioassay experiments and greenhouse rearing, lower colonization success and population growth in larger outdoor systems, and lack of establishment on Lake Gaston. Underlying mechanisms for such differences are unknown but may be due to overwintering strategies, density of the canopy at the water surface, and lowered nutritional value between the biotypes.

XIII International Symposium on Biological Control of Weeds - 2011 190 Session 4 Target and Agent Selection

Natural Enemies of Floating Marshpennywort (Hydrocotyle ranunculoides) in the Southern USA

N. E. Harms, J. F. Shearer and M. J. Grodowitz

U.S. Army Engineer Research and Development Center, 3909 Halls Ferry Rd., Vicksburg, MS 39180 USA [email protected]

Abstract

The aquatic plant, floating marshpennywort, is invasive to many areas where it has been introduced, including Australia and Europe. Because traditional methods of control are often costly, ineffective, or unsustainable, biological control is being considered as a viable alternative. The first step in initiating a successful biological control program is a survey of natural enemies of the target plant in its home range. Since it is considered native to the United States, floating marshpennywort was surveyed in the Gulf Coast States for insect herbivores and pathogens from 2007-2011. A total of ten insect and five potential pathogenic species were recovered during the surveys. Insects recovered included three weevil (Coleoptera: Curculionidae), two caterpillar (Lepidoptera), and at least one grass fly (Diptera: Chloropidae). Although many of the insect species collected are known to be generalist herbivores, several had unknown diets. The chloropid, Eugaurax floridensis Malloch, and nymphalid, Enigmogramma basigera (Walker) show the greatest promise as biological control agents because both species are able to cause damage and complete their life cycles on floating marshpennywort. Host specificity testing is needed to determine diet for both species. Potential pathogen genera collected included Alternaria, Pestalotiopsis, Colletotrichum, and Phoma. None of the pathogen species collected are likely to be host-specific, but several appear to cause considerable leaf damage to floating marshpennywort and may limit its growth and reproduction, at least to some degree. The potential for these species to be used as mycoherbicides will require further research.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 191

Can We Optimize Native-Range Survey Effort through Space and Time?

T. A. Heard, K. Bell and R. D. van Klinken

CSIRO Ecosystem Sciences, EcoSciences Precinct, 41 Boggo Rd, Dutton Park, GPO Box 2583, Brisbane, 4001, Australia [email protected]

Abstract

Native range surveys are an expensive and time consuming step in biological control projects but are crucial for determination of the arthropod and fungal fauna which provides the basis upon which agent selection can take place. A key challenge when designing and conducting native-range surveys is to decide how best to allocate survey effort through space (the species-native range distribution) and through time (at any one particular site or region). For example, should more effort be spent searching the same sites more extensively or on searching elsewhere, and when should surveying come to an end? Parkinsonia aculeata L. is a weed of Australia’s rangelands that occurs naturally over a large part of the American tropics and subtropics (USA to Argentina). Native range surveys have been conducted over many years in many parts of this range but in an uneven way with some parts intensively surveyed, some parts minimally surveyed and other parts not surveyed at all. A study commenced in 2008 which used the existing results of surveys to predict where further surveys would be most fruitful. Information used in this study included 1) biogeographic zoning of the Americas according to the insect fauna, 2) a P. aculeata climate model, 3) P. aculeata species distributions, and 4) the insect fauna collected. We used generalized dissimilarity modelling of P. aculeata fauna to provide the data for a technique called survey gap analysis. The survey gap analysis predicted areas which should yield the greatest number of new species. In addition, we used species accumulation curves to predict which existing regions further surveying would yield most new species. We discuss the potential for these and other approaches to improve the allocation of native-range survey effort.

XIII International Symposium on Biological Control of Weeds - 2011 192 Session 4 Target and Agent Selection

Potential Agent Psectrosema noxium (Diptera: Cecidomyiidae) from Kazakhstan for Saltcedar Biological Control in USA

R. Jashenko1, I. Mityaev1 and C. J. DeLoach2

1Institute of Zoology, 93 Al-Farabi Ave, Almaty, 050060, Kazakhstan [email protected] 2USDA-ARS, Grassland, Soil & Water Research Lab, 808 E. Blackland Road, Temple, TX USA 76502

Abstract

Research was conducted on eight native populations in southeastern Kazakhstan during 1995-2009. This narrow-oligophagous, univoltine, stem-galling, midge fly attacks saltcedar (Tamarix spp.) where both are native in Kazakhstan, Turkmenistan, Uzbekistan and western China. It is a candidate for biological control of saltcedars in the western United States and northern Mexico. In nature, the adults of Psectrosema noxium Marikovskij emerged from overwintered galls in late April, mated, and laid 1-30 eggs on each of several new foliage buds. The larvae emerged in 3-4 days and immediately entered the same buds. They developed slowly as the buds developed into stems, and formed colonial galls of 1-30 cells visible in early summer that became ligneous by autumn. The larvae pupated in spring in synchrony with Tamarix bud-break. Where numerous, they killed branches or even the entire Tamarix plant. Outbreak populations were seen only each 5-7 years, reduced between peaks by severe attack by four species of egg and larval parasitoids. In Kazakhstan, we observed galls of P. noxium mostly on T. ramosissima Ledeb. and sometimes on other salcedars but never on other genera of Tamaricaceae. In outdoor, uncaged tests at Almaty, females readily accepted accessions of T. ramosissima from five U.S. states, for oviposition and the larvae completed their development on it. However, close synchronization between Tamarix bud-break and adult midge emergence is critical so that oviposition can take place on just erupted buds. Eggs laid on young shoots produce only undersized galls that fall from the plant by autumn and the larvae die. Females did not oviposit in on athel, T. aphylla (L.) H.Karst., a beneficial exotic shade tree in northern Mexico and the U.S., that is not a target for biological control. Methods of culturing P. noxium in the laboratory and in field cages were developed if it is introduced into the USA.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 193

Fungi Pathogenic on Paederia spp. from Northern Thailand as Potential Biological Control Agents for Skunkvine Paederia foetida (Rubiaceae)

M. P. Ko, M. M. Ramadan and N. J. Reimer

Hawaii Department of Agriculture, Plant Pest Control Branch, 1428 S. King Street, Honolulu, HI 96814 USA [email protected]

Abstract

The skunkvinePaederia foetida L. (known locally as Maile Pilau) is one of the major invasive environmental weeds on the Hawaiian Islands. Due to its rapid spread to remote and inaccessible areas, its ability to cause substantial damage to natural ecosystems and the increasing cost of conventional control methods, its management by biological control seems to be the most appropriate means with high success potential. A survey was conducted in its native range in northern Thailand in the fall of 2010, with the aim of locating and identifying potential agents for classical biological control. Diseased tissues of Paederia species exhibiting symptoms of necrotic spots/lesions, galls and rusts were imported into the Hawaii Department of Agriculture’s Plant Pathogen Containment Facility (HDOA- PPCF) for evaluation. Most of the infected tissues were derived from Paederia pilifera Hook.f., from which several fungi were subsequently isolated. Among these fungi were the gall rust (Basidiomycota) Endophyllum paederiae (Dietel) F. Stevens & Mendiola, the Hyphomycetes Pseudocercospora paederiae Sawada ex Goh & W.H. Hsieh, and two isolates of the Coelomycetes Colletotrichum gloeosporioides (Penz.) Penz. & Sacc. Repeated attempts to establish E. paederiae on the local skunkvine P. foetida failed, indicating that this rust fungus may be too host specific to infect the local species. However, the remaining fungi P. paederiae and isolates of C. gloeosporioides were amenable to laboratory culture on standard potato dextrose agar. Separate inoculation tests on the local P. foetida plants with conidia from each fungal culture showed leaf lesions or necrotic spots, where the respective fungus could subsequently be reisolated. One of the isolates of C. gloeosporioides was relatively aggressive, causing leaf chlorosis, defoliation and even shoot tip dieback on the infected P. foetida plants. Further studies on the potentials of these fungi as biological control agents on P. foetida, such as their effects on other economic plants (host range), culture and pathogenic enhancements by environmental factors etc., are underway inside the HDOA-PPCF.

XIII International Symposium on Biological Control of Weeds - 2011 194 Session 4 Target and Agent Selection

Preliminary Surveys for Natural Enemies of the North American Native Delta Arrowhead (Sagittaria platyphylla, Alismataceae), an Invasive Species in Australia

R. M. Kwong1, J.-L. Sagliocco1, N. E. Harms2 and J. F. Shearer2

1Victorian Department of Primary Industries, PO Box 48, Frankston, Victoria, Australia, 3199 [email protected] 2U.S. Army Engineer Research and Development Center, 3909 Halls Ferry Road, Vicksburg, Mississippi, USA 39180

Abstract

The perennial, North American native delta arrowhead (Sagittaria platyphylla (Engelm.) J.G.Sm., Alismataceae) was introduced to Australia as an ornamental pond plant in the 1950s. Today, it is a serious invader of irrigation channels and natural waterways in southeastern Australia with up to $2 million being spent annually on control in the worst affected regions. Delta arrowhead is free from attack by herbivores and pathogens in Australia, where it seeds prolifically for nearly seven months of the year. Recent preliminary surveys in the southern USA identified a number of curculionids in the genus Listronotus associated with delta arrowhead, none previously reported on this host-plant. These weevils and a chloropid fly, Eugaurax sp. were all found to attack flowers and fruits. Further surveys are planned to complete the catalogue of natural enemy flora and fauna across the native range. Upcoming genetic studies to identify the genetic origin(s) of the Australian populations will underpin future research on the selection of biological control agents for delta arrowhead.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 195

Prospects for Biological Control of Berberis darwinii (Berberidaceae) in New Zealand: What are its Seed Predators in its Native Chilean Range?

H. Norambuena1, L. Smith2 and S. Rothmann3

1Pasaje El Acantilado Oriente 740, Temuco, Chile [email protected] 2Landcare Research New Zealand, Gerald Street, Lincoln PO Box 40, Lincoln 7640, New Zealand 3SAG - Lo Aguirre, Santiago, Chile

Abstract

A native of South America, Darwin’s barberry, Berberis darwinii Hook, (Berberidaceae) has become an invasive species in New Zealand. It has invaded many habitat types from grazed pastures to intact forests, due in part to its large reproductive capacity. Consequently, as an early step in a biocontrol solution surveys were conducted in its native range for damaging invertebrates utilizing flower buds and seeds. Sampling for potential agents was conducted at 35 sites in southern Chile between Concepción (36º57’ S.) and Chiloé (42º 52’ S.). At suitable sites flowering or fruiting plants of B. darwinii were beaten 5 times onto a beating tray and all weevil species observed were collected. The insect surveys yielded four weevil species on Darwin’s barberry. Berberidicola exaratus (Blanchard) was the most common and widely distributed seed predator. It was detected at 29 of the 35 sites. Anthonomus kuscheli Clark was the most common flower bud feeder and was detected at 13 of the 35 sites. Damage to the seeds and flower buds by these weevils is obvious. Host-testing studies of these two weevil species is continuing in Chile.

XIII International Symposium on Biological Control of Weeds - 2011 196 Session 4 Target and Agent Selection

Surveys for Potential Biological Control Agents for Pereskia aculeata: Selection of the Most Promising Potential Agents

I. D. Paterson1, M. P. Hill1, S. Neser2 and D. A. Downie1

1Department of Zoology and Entomology, Rhodes University, PO Box 94, Grahamstown, 6140, South Africa [email protected] 2ARC-Plant Protection Research Institute, Private Bag X134, Queenswood, 0212, South Africa

Abstract

New biological control agents are required for the control of Pereskia aculeata Mill. in order to reduce the weed’s density to acceptable levels. Data from eight surveys for natural enemies of P. aculeata in the native distribution were used to compile a list of insect species associated with the plant. Sixty-two sites were surveyed resulting in a list of 40 insect species associated with the plant. Six prioritization categories were used to identify the most promising potential biological control agents from the suite of insects associated with the plant. Prioritization categories were i) the presence of feeding damage, ii) insect incidence measured as the number of sites at which the insect species was present divided by the total number of sites, iii) the host range of the insect observed in the field, iv) the similarity of the climate where the insect species was found to the climate at the weed’s introduced distribution, v) the similarity of the weed genotype to the genotype on which the insect species developed in the native distribution and vi) the mode and levels of damage in the native distribution. The most promising potential biological control agents for P. aculeata identified using the various criteria of the prioritization categories are the released biological control agent, Phenrica guérini Bechyné (Chrysomelidae), two species of Curculionidae and Maracayia chlorisalis Walker (Crambidae). The method used to prioritize the most promising potential biological control agents for future research may be useful when surveying for natural enemies for use as biological control agents for other weed species.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 197

Predicting the Feasibility and Cost of Weed Biological Control

Q. Paynter1, J. Overton1, S. Fowler1, R. Hill2, S. Bellgard1 and M. Dawson1

1Landcare Research, New Zealand [email protected] 2Richard Hill & Associates, Christchurch, New Zealand

Abstract

We reviewed previous work related to selection and prioritisation of weed biological control targets and developed and tested hypotheses regarding which attributes make weeds more or less amenable to biological control. Our analyses revealed that biocontrol impacts have, on average, been greater against, biennial and perennial versus annual weeds, plants capable of vegetative reproduction versus plant that reproduce solely by seed or spores, aquatic and wetland weeds versus terrestrial weeds, and plants that are not reported to be weedy in the native range versus those which are known to be weedy in the native range. We incorporated these criteria affecting biological control success into a feasibility scoring framework that should enable practitioners to better prioritise targets for biological control in the future.

XIII International Symposium on Biological Control of Weeds - 2011 198 Session 4 Target and Agent Selection

USDA-ARS Australian Biological Control Laboratory

M. Purcell, J. Makinson, R. Zonneveld, B. Brown, D. Mira, G. Fichera, A. McKinnon and S. Raghu

CSIRO Ecosystem Sciences, USDA-ARS, Office of International Research Programs, Australian Biological Control Laboratory, GPO Box 2583, Brisbane, Queensland, Australia 4001 [email protected]

Abstract

The staff of the United States Department of Agriculture (USDA), Agricultural Research Service (ARS), Australian Biological Control Laboratory (ABCL) conduct exploration in the natural areas of Australia and Asia for insects and other organisms that feed on pest insects and plant species that are invasive in the USA. Based at the Ecosciences Precinct in Brisbane, Queensland, Australia, the ABCL is operated by the USDA-ARS Office of International Research Programs (OIRP) while personnel and facilities in Australia are provided through a co-operative agreement with the Commonwealth Scientific and Industrial Research Organization (CSIRO). Many invasive weeds in the USA such as the broad-leaved paperbark tree, Melaleuca quinquenervia (Cav.) S.T. Blake; Old World climbing fern, Lygodium microphyllum (Cav.) R.Br.; hydrilla, Hydrilla verticillata (L.f.) Royle; and Australian pine, Casuarina spp. are native to Australia. However, the native distribution of many of these weed species extends into tropical and subtropical Southeast Asia, including Indonesia, Malaysia, Thailand, Vietnam, Papua New Guinea, India and southern China, as well into the Pacific Islands. With excellent collaborators in these regions, ABCL has the capability to explore these countries to find the most promising biological control agents for these and other targets, and evaluate them under quarantine conditions in Brisbane. Research conducted at ABCL includes determination of the native range of a target, exploration for natural enemies, molecular typing of herbivores, ecology of the agents and their hosts, field host-range surveys and ultimately preliminary host-range screening and impact assessments of candidate agents. The data we gather on potential agents is combined with information about the ecology of the target where it is invasive. Our research seeks to determine what regulates the target in its native environment and evaluates all potential biological control agents particularly those that can mitigate the weed’s impact. Organisms with a narrow host range and good regulatory potential are prioritized and intensively investigated. In collaboration with US-based ARS scientists, agents are selected and shipped to the United States for further quarantine studies and possible release.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 199

Potential Biological Control Agents of Skunkvine, Paederia foetida (Rubiaceae), Recently Discovered in Thailand and Laos

M. M. Ramadan, W. T. Nagamine and R. C. Bautista

State of Hawaii Department of Agriculture, Division of Plant Industry, Plant Pest Control Branch, 1428 South King Street, Honolulu, HAWAII 96814, USA [email protected]

Abstract

The skunkvine, Paederia foetida L., also known as Maile pilau in Hawaii, is an invasive weed that smothers shrubs, trees, and native flora in dry to wet forests. It disrupts perennial crops and takes over landscaping in moist to wet areas on four Hawaiian Islands. Skunk-vine is considered a noxious weed in southern United States (Alabama and Florida) and also an aggressive weed in Brazil, New Guinea, Christmas and Mauritius Islands. Chemical control is difficult without non-target damage as the vine mixes up with desirable plants. Biological control is thought to be the most suitable option for long term management of the weed in Hawaii and Florida. Skunk-vine and most species of genus Paederia are native to tropical and subtropical Asia, from as far as India to Japan and Southeast Asia. There are no native plants in the tribe Paederieae in Hawaii and Florida and the potential for biological control looks promising. A recent survey in October-December 2010, after the rainy season in Thailand and Laos, confirmed the presence of several insect herbivores associated with P. foetida and three other Paederia species. A leaf- tying moth (Lepidoptera: Crambidae), two hawk moths (Lepidoptera: Sphingidae), a herbivorous rove beetle (Coleoptera: Staphylinidae), a chrysomelid leaf beetle (Coleoptera: Chrysomelidae), a sharpshooter leafhopper (Hemiptera: Cicadellidae), and a leaf-sucking lace bug (Hemiptera: Tingidae) were the most damaging to the vine during the survey period. The beetles are being investigated at the HDOA Insect Containment Facility as potential candidates for biological control of Maile pilau in Hawaii. Initial findings on host specificity, biology, and their potential for suppressing this weed are discussed.

XIII International Symposium on Biological Control of Weeds - 2011 200 Session 4 Target and Agent Selection

Towards Biological Control of Swallow-Worts: The Ugly, the Bad and the Good

R. Sforza1, M. Augé1, M.-C. Bon1, R. Dolgovskaya2, Y. Garnier1, M. Jeanneau1, J. Poidatz1, S. Reznik2, O. Simonot1, M. Volkovitch2 and L. R. Milbrath3

1USDA-ARS European Biological Control Laboratory, CS 90013 Montferrier sur Lez, 34988 St Gély du Fesc, France [email protected] 2Zoological Institute, Anglijskij prospect, 32, St.Petersburg 190121, Russia 3USDA Agricultural Research Service, Robert W. Holley Center for Agriculture and Health, 538 Tower Road, Ithaca, NY 14853, USA

Abstract

Native to Eurasia, swallow-worts (“the ugly:” Vincetoxicum rossicum (Kleopov) Barbarich and V. nigrum L. - Apocynaceae) have invaded forested landscapes and prevented native plant regeneration in eastern North America. We first aimed to understand where the invasive populations of both species come from, and then evaluated the impact of potential biological control agents (BCAs). The following phytophagous BCAs have been studied since 2009: Chrysochus asclepiadeus (Pallas) (Col., Chrysomelidae), Abrostola asclepiadis (Denis & Schiffermuller) and Abrostola clarissa (Staudinger) (Lep., Noctuidae). Adults of the beetle feed on leaves while larvae are root feeders, and Abrostola spp. larvae are foliage feeders. Genetically, none of the native V. nigrum populations analyzed to date possesses exactly the major multilocus genotype detected in the invasive North American populations, in contrast to V. rossicum, for which source populations of the invasion are found to be in Ukraine. We performed choice and no-choice specificity tests with French and Russian populations of C. asclepiadeus. We evaluated adult herbivory in no-choice tests on three Vincetoxicum spp., as controls, and seven Asclepias spp.: C. asclepiadeus fed on controls but also on Asclepias tuberosa L., a monarch butterfly host plant. Choice tests revealed no herbivory outside the genus Vincetoxicum. Larval herbivory in choice tests was noticed on all controls, plus A. tuberosa and Asclepias syriaca L. Similar results were obtained for both populations of C. asclepiadeus. Although C. asclepiadeus has a severe impact on swallow-worts, herbivory on several Asclepias spp. lead us to consider it a “bad” BCA. No-choice tests with larvae of Abrostola asclepiadis from France revealed that they died in 5d on all the Asclepias spp., but developed to pupa in 23d on Vincetoxicum hirundinaria Medik., in 20.6d on V. rossicum, and only reached the 3rd instar in 17.8d on V. nigrum. Similar results were obtained with Abrostola clarissa of Russian origin. Thus, data with Abrostola spp. appear promising, and we consider the two Abrostola species to have good potential as BCAs against all genotypes of swallow-worts.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 201

Genetic and Behavioral Differences among Purported Species of Trichosirocalus (Coleoptera: Curculionidae) for Biological Control of Thistles (Asteraceae: Cardueae)

A. De Biase1, S. Primerano1, S. Belvedere1, E. Colonnelli2, L. Smith and M. Cristofaro4

1Dept. of Biology and Biotechnologies “Charles Darwin”, University of Rome “La Sapienza”, Viale dell’Università 32, 00185 Rome, Italy 2c/o Entomological Museum of the University of Rome “La Sapienza”, Piazzale Valerio Massimo 1, 00185 Rome, Italy 3USDA-ARS, 800 Buchanan Street, Albany, CA 94710, USA [email protected] 4ENEA C.R. Casaccia, UTAGRI-ECO, Via Anguillarese, 301 00123 S. Maria di Galeria Rome, Italy

Abstract

Trichosirocalus horridus (Panzer) was introduced to North America, New Zealand and Australia for biological control of Carduus nutans L. Since then two more species of Trichosirocalus have been described (Alonso-Zarazaga and Sánchez-Ruiz. 2002. Aust. J. Entomol. 41: 199-208), and the three species are thought to have different host plant associations: T. horridus on Cirsium vulgare (Savi) Ten. and possibly on other Carduineae, T. mortadelo Alonso-Zarazaga & Sanchez-Ruiz on C. nutans, and T. briesei Alonso- Zarazaga & Sanchez-Ruiz on Onopordum spp. This raises the question of which species were previously released for biological control of C. nutans. Subsequent studies by Groenteman et al. (2009, XII International Symposium on Biological Control of Weeds, pp. 145-149.) raises uncertainty about which species are in New Zealand and whether T. mortadelo was a valid species. Trichosirocalus briesei was introduced to Australia to control Onopordum spp. and is being evaluated for introduction to North America. We analyzed part of the mtDNA cytochrome oxidase I (COI) gene sequence of adult specimens representing the three species collected in Spain, Italy, USA, New Zealand and Australia. All specimens morphologically identified as T. briesei formed one clade that was clearly distinct from all the other specimens. The COI sequences for specimens of T. horridus and T. mortadelo were intermixed within the same clade, suggesting that they represent one heterogeneous species. Furthermore, the morphological characters attributed to T. mortadelo are of little significance to really isolate two different species, so that we combine them under the name of T. horridus. In laboratory choice experiments, specimens from Spain identified as T. briesei preferred Onopordum acanthium L. to Carduus or Cirsium spp., whereas those from North America, identified as T. horridus preferred Carduus spp. but also attacked Cirsium spp.

XIII International Symposium on Biological Control of Weeds - 2011 202 Session 4 Target and Agent Selection

Survey of Dispersal and Genetic Variability of Tectococcus ovatus (Heteroptera: Eriococcidae) in the Regions of Natural Occurrence of Psidium cattleianum (Myrtaceae)

L. E. Ranuci1, T. Johnson2 and M. D. Vitorino1,3

1Masters Program in Environmental Engineering, Blumenau Regional University, Brazil 2USDA Forest Service, Volcano, Hawaii, USA 3Forest Engineering Department, Blumenau Regional University, Brazil [email protected]

Abstract

The species Psidium cattleianum L. is considered one of the greatest threats to the ecosystem and biodiversity of the islands of Hawaii. Seeking to control its dissemination, techniques of biological control were used. Among the various species studied, as a biological agent control, Tectococcus ovatus Hempel showed a higher level of specificity. This work had as aim to verify the existence of genetic variability among and inside the different populations of T. ovatus, using the technique of PCR-RAPD. The analyses were made from females collected in the states of Rio de Janeiro, Paraná, Santa Catarina and Rio Grande do Sul in Brazil. From the eight initiators of PCR-RAPD tested, four were used in the analyses, revealing monomorphic and polymorphic markers with a variable frequency, to the individuals of one place as well as to the individuals of different places. Through the analysis of the grouping of molecular characterization it was possible to verify a formation of two distinctive groups A and B presenting a genetic variability/variable of 44%. The results obtained through the analysis of markers RAPD were useful in the verifying of variation and provided safe information about the levels of variability and similarity amongst and inside the different populations of T. ovatus.

XIII International Symposium on Biological Control of Weeds - 2011 Session 4 Target and Agent Selection 203

Arundo donax – Giant Reed

P. M o r a n 1, J. Adamczyk1, A. Racelis1, A. Kirk2, K. Hoelmer2, J. Everitt1, C. Yang1, M. Ciomperlik3, T. Roland3, R. Penk3, K. Jones3, D. Spencer4, A. Pepper5, J. Manhart5, D. Tarin5, G. Moore5, R. Lacewell6, E. Rister6, A. Sturdivant6, B. Contreras Arquieta7, M. Martínez Jiménez8, M. Marcos9, E. Cortés Mendoza9, E. Chilton10, L. Gilbert11, T. Vaughn12, A. Rubio12, R. Summy13, D. Foley14, C. Foley15 and F. Nibling16

1United States Dept. of Agriculture, Agricultural Research Service, Kika de la Garza Subtropical Agricultural Research Center, Weslaco, TX, USA [email protected] 2USDA-ARS, European Biological Control Laboratory, Montpelier, France 3USDA-APHIS, Edinburg, TX, USA 4Invasive & Exotic Research Unit, Davis, CA, USA 5Texas A&M Univ., Dept. of Biology, College Station, TX; 5Texas A&M Dept. of Rangeland Ecology 6Texas Agrilife Research & Extension, College Station, TX, USA 7Pronatura Norestre, Monterrey, Mexico 8Instituto Mexicano de Tecnología del Agua, Jiutepec, Mexico 9Universidad de Alicante, Spain 10Texas Parks & Wildlife, Austin, TX, USA 11Univ. of Texas, Austin, TX, USA 12Texas A&M International, Laredo, TX, USA 13Univ. of Texas – Pan American, Edinburg, TX, USA 14Sul Ross Univ., Del Rio, TX, USA 15Southwest Texas College, Del Rio, TX, USA 16Bureau of Reclamation, Denver, CO, USA

Abstract

Arundo donax L., giant reed, carrizo cane, is an exotic and invasive weed of riparian habitats in the southwestern U.S. and northern Mexico. Giant reed dominates these habitats, which leads to: loss of biodiversity; stream bank erosion; damage to bridges; increased costs for chemical or mechanical control along irrigation canals and transportation corridors; and impedes access for law enforcement personnel. Most importantly this invasive weed competes for water resources in an arid region where these resources are critical to the environment, agriculture and urban users. Biological control using insects from the native range of giant reed may be the best option for long-term management. A. donax is a good target for biological control because it has no close relatives in North or South America, and several of the plant feeding insects from Mediterranean Europe and known to only feed on A. donax.

XIII International Symposium on Biological Control of Weeds - 2011 204 Session 4 Target and Agent Selection

Foreign Exploration for Biological Control Agents of Giant Reed, Arundo donax

J. A. Goolsby1, P. J. Moran1 and R. Carruthers2

1United States Dept. of Agriculture, Agricultural Research Service, Kika de la Garza Subtropical Agricultural Research Center, Weslaco, TX, USA [email protected] 2United States Dept. of Agriculture, Agricultural Research Service, Exotic and Invasive Weed Research Unit, Albany, CA, USA

Abstract

Collections of insect-infested Arundo donax L. have been made 2000 - 2011 by scientists at the USDA/ARS EBCL facility in Montpellier, France. Pieces of A. donax stem, leaves and rhizomes were placed in double wrapped paper bags and kept in tight mesh covered boxes in the EBCL insect quarantine. Emergence of insects was noted and specimens sent to relevant authorities for identification. Samples were taken at the same time for genetic characterization of A. donax. Surveys have been made in appropriate areas of Croatia, Bulgaria, Slovenia, France, Spain, Portugal, Canary Isles, Italy, Greece, Crete, Turkey, Morocco, Tunisia, Egypt, South Africa, Namibia, Kenya and Australia. Surveys were in fall and spring to cover the most important growth periods. Site details, locality, altitude, GPS position were recorded. Giant reed rhizomes were unearthed and dissected at some sites for natural enemies. Lengths of rhizome and cut stems and leaves were placed in moisture absorbent bags, cooled, and returned to the EBCL quarantine for emergence. Quadrats (50x50cm) of A. donax have been sampled from 6 stands each week for 15 weeks (starting May 5 2003 in 2004 and 2005) in the Montpellier and Perpignan areas of southern France. All A. donax within the quadrats was cut, taken back to the laboratory, examined, dissected and documented. Organisms found were where possible reared and adults passed on to appropriate taxonomists. The arthropod herbivores collected from A. donax were (in order of most to least common) Tetramesa romana Walker (Hymenotpera: Eurytomidae); Rhizaspidiotus donacis (Leonardi) (Hemiptera: Diaspididae); Cryptonevra spp. (Diptera: Chloropidae); Lasioptera donacis Coutin (Diptera: Ceccidomyidae); Cerodontha phragmitidis Nowakowski (Diptera: Agromyzidae); Melanaphis donacis (Passerini) (Hemiptera: Aphidae); Aclerda berlesii Buffa (Hemiptera: Aclerdidae); Siteroptes sp. (Acarina: Pyemotidae); and Hypogaea sp. (Hemiptera: Aphididae). Only the first four species were found to be sufficiently host specific to warrant further host range testing.

XIII International Symposium on Biological Control of Weeds - 2011